redict/tests/integration/aof-multi-part.tcl

1333 lines
52 KiB
Tcl
Raw Normal View History

Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
source tests/support/aofmanifest.tcl
set defaults {appendonly {yes} appendfilename {appendonly.aof} appenddirname {appendonlydir} auto-aof-rewrite-percentage {0}}
set server_path [tmpdir server.multi.aof]
set aof_dirname "appendonlydir"
set aof_basename "appendonly.aof"
set aof_dirpath "$server_path/$aof_dirname"
set aof_base1_file "$server_path/$aof_dirname/${aof_basename}.1$::base_aof_sufix$::aof_format_suffix"
set aof_base2_file "$server_path/$aof_dirname/${aof_basename}.2$::base_aof_sufix$::aof_format_suffix"
set aof_incr1_file "$server_path/$aof_dirname/${aof_basename}.1$::incr_aof_sufix$::aof_format_suffix"
set aof_incr2_file "$server_path/$aof_dirname/${aof_basename}.2$::incr_aof_sufix$::aof_format_suffix"
set aof_incr3_file "$server_path/$aof_dirname/${aof_basename}.3$::incr_aof_sufix$::aof_format_suffix"
set aof_manifest_file "$server_path/$aof_dirname/${aof_basename}$::manifest_suffix"
set aof_old_name_old_path "$server_path/$aof_basename"
set aof_old_name_new_path "$aof_dirpath/$aof_basename"
set aof_old_name_old_path2 "$server_path/${aof_basename}2"
set aof_manifest_file2 "$server_path/$aof_dirname/${aof_basename}2$::manifest_suffix"
tags {"external:skip"} {
# Test Part 1
# In order to test the loading logic of redis under different combinations of manifest and AOF.
# We will manually construct the manifest file and AOF, and then start redis to verify whether
# the redis behavior is as expected.
test {Multi Part AOF can't load data when some file missing} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr2_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
append_to_manifest "file appendonly.aof.2.incr.aof seq 2 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "appendonly.aof.1.incr.aof .*No such file or directory"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the sequence not increase monotonically} {
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr2_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.2.incr.aof seq 2 type i\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "Found a non-monotonic sequence number"]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when there are blank lines in the manifest file} {
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr3_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
append_to_manifest "\n"
append_to_manifest "file appendonly.aof.3.incr.aof seq 3 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "Invalid AOF manifest file format"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when there is a duplicate base file} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_base2_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.2.base.aof seq 2 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "Found duplicate base file information"]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest format is wrong (type unknown)} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type x\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "Unknown AOF file type"]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest format is wrong (missing key)} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "filx appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 2 [count_message_lines $server_path/stdout "Invalid AOF manifest file format"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest format is wrong (line too short)} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 3 [count_message_lines $server_path/stdout "Invalid AOF manifest file format"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest format is wrong (line too long)} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "The AOF manifest file contains too long line"]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest format is wrong (odd parameter)} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i newkey\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 4 [count_message_lines $server_path/stdout "Invalid AOF manifest file format"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can't load data when the manifest file is empty} {
create_aof_manifest $aof_dirpath $aof_manifest_file {
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "Found an empty AOF manifest"]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can start when no aof and no manifest} {
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
assert_equal OK [$client set k1 v1]
assert_equal v1 [$client get k1]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can start when we have en empty AOF dir} {
create_aof_dir $aof_dirpath
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
}
test {Multi Part AOF can load data discontinuously increasing sequence} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof $aof_dirpath $aof_incr3_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
append_to_manifest "file appendonly.aof.3.incr.aof seq 3 type i\n"
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal v1 [$client get k1]
assert_equal v2 [$client get k2]
assert_equal v3 [$client get k3]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can load data when manifest add new k-v} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
append_to_aof [formatCommand set k2 v2]
}
create_aof $aof_dirpath $aof_incr3_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b newkey newvalue\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
append_to_manifest "file appendonly.aof.3.incr.aof seq 3 type i\n"
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal v1 [$client get k1]
assert_equal v2 [$client get k2]
assert_equal v3 [$client get k3]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can load data when some AOFs are empty} {
create_aof $aof_dirpath $aof_base1_file {
append_to_aof [formatCommand set k1 v1]
}
create_aof $aof_dirpath $aof_incr1_file {
}
create_aof $aof_dirpath $aof_incr3_file {
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof.1.base.aof seq 1 type b\n"
append_to_manifest "file appendonly.aof.1.incr.aof seq 1 type i\n"
append_to_manifest "file appendonly.aof.3.incr.aof seq 3 type i\n"
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal v1 [$client get k1]
assert_equal "" [$client get k2]
assert_equal v3 [$client get k3]
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can load data from old version redis (rdb preamble no)} {
create_aof $server_path $aof_old_name_old_path {
append_to_aof [formatCommand set k1 v1]
append_to_aof [formatCommand set k2 v2]
append_to_aof [formatCommand set k3 v3]
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal v1 [$client get k1]
assert_equal v2 [$client get k2]
assert_equal v3 [$client get k3]
assert_equal 0 [check_file_exist $server_path $aof_basename]
assert_equal 1 [check_file_exist $aof_dirpath $aof_basename]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
assert_equal OK [$client set k4 v4]
$client bgrewriteaof
waitForBgrewriteaof $client
assert_equal OK [$client set k5 v5]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
}
set d1 [$client debug digest]
$client debug loadaof
set d2 [$client debug digest]
assert {$d1 eq $d2}
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can load data from old version redis (rdb preamble yes)} {
exec cp tests/assets/rdb-preamble.aof $aof_old_name_old_path
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
# k1 k2 in rdb header and k3 in AOF tail
assert_equal v1 [$client get k1]
assert_equal v2 [$client get k2]
assert_equal v3 [$client get k3]
assert_equal 0 [check_file_exist $server_path $aof_basename]
assert_equal 1 [check_file_exist $aof_dirpath $aof_basename]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
assert_equal OK [$client set k4 v4]
$client bgrewriteaof
waitForBgrewriteaof $client
assert_equal OK [$client set k5 v5]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
}
set d1 [$client debug digest]
$client debug loadaof
set d2 [$client debug digest]
assert {$d1 eq $d2}
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can continue the upgrade from the interrupted upgrade state} {
create_aof $server_path $aof_old_name_old_path {
append_to_aof [formatCommand set k1 v1]
append_to_aof [formatCommand set k2 v2]
append_to_aof [formatCommand set k3 v3]
}
# Create a layout of an interrupted upgrade (interrupted before the rename).
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof seq 1 type b\n"
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal v1 [$client get k1]
assert_equal v2 [$client get k2]
assert_equal v3 [$client get k3]
assert_equal 0 [check_file_exist $server_path $aof_basename]
assert_equal 1 [check_file_exist $aof_dirpath $aof_basename]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can be loaded correctly when both server dir and aof dir contain old AOF} {
create_aof $server_path $aof_old_name_old_path {
append_to_aof [formatCommand set k1 v1]
append_to_aof [formatCommand set k2 v2]
append_to_aof [formatCommand set k3 v3]
}
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof seq 1 type b\n"
}
create_aof $aof_dirpath $aof_old_name_new_path {
append_to_aof [formatCommand set k4 v4]
append_to_aof [formatCommand set k5 v5]
append_to_aof [formatCommand set k6 v6]
}
start_server_aof [list dir $server_path] {
assert_equal 1 [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set client [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_done_loading $client
assert_equal 0 [$client exists k1]
assert_equal 0 [$client exists k2]
assert_equal 0 [$client exists k3]
assert_equal v4 [$client get k4]
assert_equal v5 [$client get k5]
assert_equal v6 [$client get k6]
assert_equal 1 [check_file_exist $server_path $aof_basename]
assert_equal 1 [check_file_exist $aof_dirpath $aof_basename]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
}
clean_aof_persistence $aof_dirpath
catch {exec rm -rf $aof_old_name_old_path}
}
test {Multi Part AOF can't load data when the manifest contains the old AOF file name but the file does not exist in server dir and aof dir} {
create_aof_manifest $aof_dirpath $aof_manifest_file {
append_to_manifest "file appendonly.aof seq 1 type b\n"
}
start_server_aof_ex [list dir $server_path] [list wait_ready false] {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
wait_for_condition 100 50 {
! [is_alive [srv pid]]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
} else {
fail "AOF loading didn't fail"
}
assert_equal 1 [count_message_lines $server_path/stdout "appendonly.aof .*No such file or directory"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can upgrade when when two redis share the same server dir} {
create_aof $server_path $aof_old_name_old_path {
append_to_aof [formatCommand set k1 v1]
append_to_aof [formatCommand set k2 v2]
append_to_aof [formatCommand set k3 v3]
}
create_aof $server_path $aof_old_name_old_path2 {
append_to_aof [formatCommand set k4 v4]
append_to_aof [formatCommand set k5 v5]
append_to_aof [formatCommand set k6 v6]
}
start_server_aof [list dir $server_path] {
set redis1 [redis [srv host] [srv port] 0 $::tls]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
start_server [list overrides [list dir $server_path appendonly yes appendfilename appendonly.aof2]] {
set redis2 [redis [srv host] [srv port] 0 $::tls]
test "Multi Part AOF can upgrade when when two redis share the same server dir (redis1)" {
wait_done_loading $redis1
assert_equal v1 [$redis1 get k1]
assert_equal v2 [$redis1 get k2]
assert_equal v3 [$redis1 get k3]
assert_equal 0 [$redis1 exists k4]
assert_equal 0 [$redis1 exists k5]
assert_equal 0 [$redis1 exists k6]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
$redis1 bgrewriteaof
waitForBgrewriteaof $redis1
assert_equal OK [$redis1 set k v]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
}
set d1 [$redis1 debug digest]
$redis1 debug loadaof
set d2 [$redis1 debug digest]
assert {$d1 eq $d2}
}
test "Multi Part AOF can upgrade when when two redis share the same server dir (redis2)" {
wait_done_loading $redis2
assert_equal 0 [$redis2 exists k1]
assert_equal 0 [$redis2 exists k2]
assert_equal 0 [$redis2 exists k3]
assert_equal v4 [$redis2 get k4]
assert_equal v5 [$redis2 get k5]
assert_equal v6 [$redis2 get k6]
assert_aof_manifest_content $aof_manifest_file2 {
{file appendonly.aof2 seq 1 type b}
{file appendonly.aof2.1.incr.aof seq 1 type i}
}
$redis2 bgrewriteaof
waitForBgrewriteaof $redis2
assert_equal OK [$redis2 set k v]
assert_aof_manifest_content $aof_manifest_file2 {
{file appendonly.aof2.2.base.rdb seq 2 type b}
{file appendonly.aof2.2.incr.aof seq 2 type i}
}
set d1 [$redis2 debug digest]
$redis2 debug loadaof
set d2 [$redis2 debug digest]
assert {$d1 eq $d2}
}
}
}
}
test {Multi Part AOF can handle appendfilename contains whitespaces} {
start_server [list overrides [list appendonly yes appendfilename "\" file seq \\n\\n.aof \""]] {
set dir [get_redis_dir]
set aof_manifest_name [format "%s/%s/%s%s" $dir "appendonlydir" " file seq \n\n.aof " $::manifest_suffix]
set redis [redis [srv host] [srv port] 0 $::tls]
assert_equal OK [$redis set k1 v1]
$redis bgrewriteaof
waitForBgrewriteaof $redis
assert_aof_manifest_content $aof_manifest_name {
{file " file seq \n\n.aof .2.base.rdb" seq 2 type b}
{file " file seq \n\n.aof .2.incr.aof" seq 2 type i}
}
set d1 [$redis debug digest]
$redis debug loadaof
set d2 [$redis debug digest]
assert {$d1 eq $d2}
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can create BASE (RDB format) when redis starts from empty} {
start_server_aof [list dir $server_path] {
set client [redis [srv host] [srv port] 0 $::tls]
wait_done_loading $client
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::base_aof_sufix}${::rdb_format_suffix}"]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.1.base.rdb seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
$client set foo behavior
set d1 [$client debug digest]
$client debug loadaof
set d2 [$client debug digest]
assert {$d1 eq $d2}
}
clean_aof_persistence $aof_dirpath
}
test {Multi Part AOF can create BASE (AOF format) when redis starts from empty} {
start_server_aof [list dir $server_path aof-use-rdb-preamble no] {
set client [redis [srv host] [srv port] 0 $::tls]
wait_done_loading $client
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::base_aof_sufix}${::aof_format_suffix}"]
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.1.base.aof seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
$client set foo behavior
set d1 [$client debug digest]
$client debug loadaof
set d2 [$client debug digest]
assert {$d1 eq $d2}
}
clean_aof_persistence $aof_dirpath
}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
# Test Part 2
#
# To test whether the AOFRW behaves as expected during the redis run.
# We will start redis first, then perform pressure writing, enable and disable AOF, and manually
# and automatically run bgrewrite and other actions, to test whether the correct AOF file is created,
# whether the correct manifest is generated, whether the data can be reload correctly under continuous
# writing pressure, etc.
start_server {tags {"Multi Part AOF"} overrides {aof-use-rdb-preamble {yes} appendonly {no} save {}}} {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
set dir [get_redis_dir]
set aof_basename "appendonly.aof"
set aof_dirname "appendonlydir"
set aof_dirpath "$dir/$aof_dirname"
set aof_manifest_name "$aof_basename$::manifest_suffix"
set aof_manifest_file "$dir/$aof_dirname/$aof_manifest_name"
set master [srv 0 client]
set master_host [srv 0 host]
set master_port [srv 0 port]
catch {exec rm -rf $aof_manifest_file}
test "Make sure aof manifest $aof_manifest_name not in aof directory" {
assert_equal 0 [file exists $aof_manifest_file]
}
test "AOF enable will create manifest file" {
r config set appendonly yes ; # Will create manifest and new INCR aof
r config set auto-aof-rewrite-percentage 0 ; # Disable auto-rewrite.
waitForBgrewriteaof r
# Start write load
set load_handle0 [start_write_load $master_host $master_port 10]
wait_for_condition 50 100 {
[r dbsize] > 0
} else {
fail "No write load detected."
}
# First AOFRW done
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.1.base.rdb seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
# Check we really have these files
assert_equal 1 [check_file_exist $aof_dirpath $aof_manifest_name]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::incr_aof_sufix}${::aof_format_suffix}"]
r bgrewriteaof
waitForBgrewriteaof r
# The second AOFRW done
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
}
assert_equal 1 [check_file_exist $aof_dirpath $aof_manifest_name]
# Wait bio delete history
wait_for_condition 1000 10 {
[check_file_exist $aof_dirpath "${aof_basename}.1${::base_aof_sufix}${::rdb_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.1${::incr_aof_sufix}${::aof_format_suffix}"] == 0
} else {
fail "Failed to delete history AOF"
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.2${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.2${::incr_aof_sufix}${::aof_format_suffix}"]
stop_write_load $load_handle0
wait_load_handlers_disconnected
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
}
test "AOF multiple rewrite failures will open multiple INCR AOFs" {
# Start write load
r config set rdb-key-save-delay 10000000
set orig_size [r dbsize]
set load_handle0 [start_write_load $master_host $master_port 10]
wait_for_condition 50 100 {
[r dbsize] > $orig_size
} else {
fail "No write load detected."
}
# Let AOFRW fail three times
r bgrewriteaof
set pid1 [get_child_pid 0]
catch {exec kill -9 $pid1}
waitForBgrewriteaof r
r bgrewriteaof
set pid2 [get_child_pid 0]
catch {exec kill -9 $pid2}
waitForBgrewriteaof r
r bgrewriteaof
set pid3 [get_child_pid 0]
catch {exec kill -9 $pid3}
waitForBgrewriteaof r
assert_equal 0 [check_file_exist $dir "temp-rewriteaof-bg-$pid1.aof"]
assert_equal 0 [check_file_exist $dir "temp-rewriteaof-bg-$pid2.aof"]
assert_equal 0 [check_file_exist $dir "temp-rewriteaof-bg-$pid3.aof"]
# We will have four INCR AOFs
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
{file appendonly.aof.3.incr.aof seq 3 type i}
{file appendonly.aof.4.incr.aof seq 4 type i}
{file appendonly.aof.5.incr.aof seq 5 type i}
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.2${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.2${::incr_aof_sufix}${::aof_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.3${::incr_aof_sufix}${::aof_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.4${::incr_aof_sufix}${::aof_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.5${::incr_aof_sufix}${::aof_format_suffix}"]
stop_write_load $load_handle0
wait_load_handlers_disconnected
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
r config set rdb-key-save-delay 0
catch {exec kill -9 [get_child_pid 0]}
wait_for_condition 1000 10 {
[s rdb_bgsave_in_progress] eq 0
} else {
fail "bgsave did not stop in time"
}
# AOFRW success
r bgrewriteaof
waitForBgrewriteaof r
# All previous INCR AOFs have become history
# and have be deleted
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.3.base.rdb seq 3 type b}
{file appendonly.aof.6.incr.aof seq 6 type i}
}
# Wait bio delete history
wait_for_condition 1000 10 {
[check_file_exist $aof_dirpath "${aof_basename}.2${::base_aof_sufix}${::rdb_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.2${::incr_aof_sufix}${::aof_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.3${::incr_aof_sufix}${::aof_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.4${::incr_aof_sufix}${::aof_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.5${::incr_aof_sufix}${::aof_format_suffix}"] == 0
} else {
fail "Failed to delete history AOF"
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.3${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.6${::incr_aof_sufix}${::aof_format_suffix}"]
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
}
test "AOF rewrite doesn't open new aof when AOF turn off" {
r config set appendonly no
r bgrewriteaof
waitForBgrewriteaof r
# We only have BASE AOF, no INCR AOF
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.4.base.rdb seq 4 type b}
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.4${::base_aof_sufix}${::rdb_format_suffix}"]
wait_for_condition 1000 10 {
[check_file_exist $aof_dirpath "${aof_basename}.6${::incr_aof_sufix}${::aof_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.7${::incr_aof_sufix}${::aof_format_suffix}"] == 0
} else {
fail "Failed to delete history AOF"
}
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
# Turn on AOF again
r config set appendonly yes
waitForBgrewriteaof r
# A new INCR AOF was created
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.5.base.rdb seq 5 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
# Wait bio delete history
wait_for_condition 1000 10 {
[check_file_exist $aof_dirpath "${aof_basename}.4${::base_aof_sufix}${::rdb_format_suffix}"] == 0
} else {
fail "Failed to delete history AOF"
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.5${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::incr_aof_sufix}${::aof_format_suffix}"]
}
test "AOF enable/disable auto gc" {
r config set aof-disable-auto-gc yes
r bgrewriteaof
waitForBgrewriteaof r
r bgrewriteaof
waitForBgrewriteaof r
# We can see four history AOFs (Evolved from two BASE and two INCR)
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.7.base.rdb seq 7 type b}
{file appendonly.aof.2.incr.aof seq 2 type h}
{file appendonly.aof.6.base.rdb seq 6 type h}
{file appendonly.aof.1.incr.aof seq 1 type h}
{file appendonly.aof.5.base.rdb seq 5 type h}
{file appendonly.aof.3.incr.aof seq 3 type i}
}
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.5${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.6${::base_aof_sufix}${::rdb_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.1${::incr_aof_sufix}${::aof_format_suffix}"]
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.2${::incr_aof_sufix}${::aof_format_suffix}"]
r config set aof-disable-auto-gc no
# Auto gc success
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.7.base.rdb seq 7 type b}
{file appendonly.aof.3.incr.aof seq 3 type i}
}
# wait bio delete history
wait_for_condition 1000 10 {
[check_file_exist $aof_dirpath "${aof_basename}.5${::base_aof_sufix}${::rdb_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.6${::base_aof_sufix}${::rdb_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.1${::incr_aof_sufix}${::aof_format_suffix}"] == 0 &&
[check_file_exist $aof_dirpath "${aof_basename}.2${::incr_aof_sufix}${::aof_format_suffix}"] == 0
} else {
fail "Failed to delete history AOF"
}
}
test "AOF can produce consecutive sequence number after reload" {
# Current manifest, BASE seq 7 and INCR seq 3
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.7.base.rdb seq 7 type b}
{file appendonly.aof.3.incr.aof seq 3 type i}
}
r debug loadaof
# Trigger AOFRW
r bgrewriteaof
waitForBgrewriteaof r
# Now BASE seq is 8 and INCR seq is 4
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.8.base.rdb seq 8 type b}
{file appendonly.aof.4.incr.aof seq 4 type i}
}
}
test "AOF enable during BGSAVE will not write data util AOFRW finish" {
r config set appendonly no
r config set save ""
r config set rdb-key-save-delay 10000000
r set k1 v1
r bgsave
wait_for_condition 1000 10 {
[s rdb_bgsave_in_progress] eq 1
} else {
fail "bgsave did not start in time"
}
# Make server.aof_rewrite_scheduled = 1
r config set appendonly yes
assert_equal [s aof_rewrite_scheduled] 1
# Not open new INCR aof
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.8.base.rdb seq 8 type b}
{file appendonly.aof.4.incr.aof seq 4 type i}
}
r set k2 v2
r debug loadaof
# Both k1 and k2 lost
assert_equal 0 [r exists k1]
assert_equal 0 [r exists k2]
set total_forks [s total_forks]
assert_equal [s rdb_bgsave_in_progress] 1
r config set rdb-key-save-delay 0
catch {exec kill -9 [get_child_pid 0]}
wait_for_condition 1000 10 {
[s rdb_bgsave_in_progress] eq 0
} else {
fail "bgsave did not stop in time"
}
# Make sure AOFRW was scheduled
wait_for_condition 1000 10 {
[s total_forks] == [expr $total_forks + 1]
} else {
fail "aof rewrite did not scheduled"
}
waitForBgrewriteaof r
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.9.base.rdb seq 9 type b}
{file appendonly.aof.5.incr.aof seq 5 type i}
}
r set k3 v3
r debug loadaof
assert_equal v3 [r get k3]
}
test "AOF will trigger limit when AOFRW fails many times" {
# Clear all data and trigger a successful AOFRW, so we can let
# server.aof_current_size equal to 0
r flushall
r bgrewriteaof
waitForBgrewriteaof r
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
r config set rdb-key-save-delay 10000000
# Let us trigger AOFRW easily
r config set auto-aof-rewrite-percentage 1
r config set auto-aof-rewrite-min-size 1kb
# Set a key so that AOFRW can be delayed
r set k v
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
# Let AOFRW fail 3 times, this will trigger AOFRW limit
r bgrewriteaof
catch {exec kill -9 [get_child_pid 0]}
waitForBgrewriteaof r
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
r bgrewriteaof
catch {exec kill -9 [get_child_pid 0]}
waitForBgrewriteaof r
r bgrewriteaof
catch {exec kill -9 [get_child_pid 0]}
waitForBgrewriteaof r
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.10.base.rdb seq 10 type b}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
{file appendonly.aof.6.incr.aof seq 6 type i}
{file appendonly.aof.7.incr.aof seq 7 type i}
{file appendonly.aof.8.incr.aof seq 8 type i}
{file appendonly.aof.9.incr.aof seq 9 type i}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
# Write 1KB data to trigger AOFRW
r set x [string repeat x 1024]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
# Make sure we have limit log
wait_for_condition 1000 50 {
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
[count_log_message 0 "triggered the limit"] == 1
} else {
fail "aof rewrite did not trigger limit"
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
assert_equal [status r aof_rewrite_in_progress] 0
# No new INCR AOF be created
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.10.base.rdb seq 10 type b}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
{file appendonly.aof.6.incr.aof seq 6 type i}
{file appendonly.aof.7.incr.aof seq 7 type i}
{file appendonly.aof.8.incr.aof seq 8 type i}
{file appendonly.aof.9.incr.aof seq 9 type i}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
# Turn off auto rewrite
r config set auto-aof-rewrite-percentage 0
r config set rdb-key-save-delay 0
catch {exec kill -9 [get_child_pid 0]}
wait_for_condition 1000 10 {
[s aof_rewrite_in_progress] eq 0
} else {
fail "aof rewrite did not stop in time"
}
# We can still manually execute AOFRW immediately
r bgrewriteaof
waitForBgrewriteaof r
# Can create New INCR AOF
assert_equal 1 [check_file_exist $aof_dirpath "${aof_basename}.10${::incr_aof_sufix}${::aof_format_suffix}"]
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.11.base.rdb seq 11 type b}
{file appendonly.aof.10.incr.aof seq 10 type i}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
}
start_server {overrides {aof-use-rdb-preamble {yes} appendonly {no} save {}}} {
set dir [get_redis_dir]
set aof_basename "appendonly.aof"
set aof_dirname "appendonlydir"
set aof_dirpath "$dir/$aof_dirname"
set aof_manifest_name "$aof_basename$::manifest_suffix"
set aof_manifest_file "$dir/$aof_dirname/$aof_manifest_name"
set master [srv 0 client]
set master_host [srv 0 host]
set master_port [srv 0 port]
test "AOF will open a temporary INCR AOF to accumulate data until the first AOFRW success when AOF is dynamically enabled" {
r config set save ""
# Increase AOFRW execution time to give us enough time to kill it
r config set rdb-key-save-delay 10000000
# Start write load
set load_handle0 [start_write_load $master_host $master_port 10]
wait_for_condition 50 100 {
[r dbsize] > 0
} else {
fail "No write load detected."
}
# Enable AOF will trigger an initialized AOFRW
r config set appendonly yes
# Let AOFRW fail
assert_equal 1 [s aof_rewrite_in_progress]
set pid1 [get_child_pid 0]
catch {exec kill -9 $pid1}
# Wait for AOFRW to exit and delete temp incr aof
wait_for_condition 1000 100 {
[count_log_message 0 "Removing the temp incr aof file"] == 1
} else {
fail "temp aof did not delete"
}
# Make sure manifest file is not created
assert_equal 0 [check_file_exist $aof_dirpath $aof_manifest_name]
# Make sure BASE AOF is not created
assert_equal 0 [check_file_exist $aof_dirpath "${aof_basename}.1${::base_aof_sufix}${::rdb_format_suffix}"]
# Make sure the next AOFRW has started
wait_for_condition 1000 50 {
[s aof_rewrite_in_progress] == 1
} else {
fail "aof rewrite did not scheduled"
}
# Do a successful AOFRW
set total_forks [s total_forks]
r config set rdb-key-save-delay 0
catch {exec kill -9 [get_child_pid 0]}
# Make sure the next AOFRW has started
wait_for_condition 1000 10 {
[s total_forks] == [expr $total_forks + 1]
} else {
fail "aof rewrite did not scheduled"
}
waitForBgrewriteaof r
assert_equal 2 [count_log_message 0 "Removing the temp incr aof file"]
# BASE and INCR AOF are successfully created
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.1.base.rdb seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
stop_write_load $load_handle0
wait_load_handlers_disconnected
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
# Dynamic disable AOF again
r config set appendonly no
# Disabling AOF does not delete previous AOF files
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
assert_equal 0 [s rdb_changes_since_last_save]
r config set rdb-key-save-delay 10000000
set load_handle0 [start_write_load $master_host $master_port 10]
wait_for_condition 50 100 {
[s rdb_changes_since_last_save] > 0
} else {
fail "No write load detected."
}
# Re-enable AOF
r config set appendonly yes
# Let AOFRW fail
assert_equal 1 [s aof_rewrite_in_progress]
set pid1 [get_child_pid 0]
catch {exec kill -9 $pid1}
# Wait for AOFRW to exit and delete temp incr aof
wait_for_condition 1000 100 {
[count_log_message 0 "Removing the temp incr aof file"] == 3
} else {
fail "temp aof did not delete 3 times"
}
# Make sure no new incr AOF was created
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.1.base.rdb seq 1 type b}
{file appendonly.aof.1.incr.aof seq 1 type i}
}
# Make sure the next AOFRW has started
wait_for_condition 1000 50 {
[s aof_rewrite_in_progress] == 1
} else {
fail "aof rewrite did not scheduled"
}
# Do a successful AOFRW
set total_forks [s total_forks]
r config set rdb-key-save-delay 0
catch {exec kill -9 [get_child_pid 0]}
wait_for_condition 1000 10 {
[s total_forks] == [expr $total_forks + 1]
} else {
fail "aof rewrite did not scheduled"
}
waitForBgrewriteaof r
assert_equal 4 [count_log_message 0 "Removing the temp incr aof file"]
# New BASE and INCR AOF are successfully created
assert_aof_manifest_content $aof_manifest_file {
{file appendonly.aof.2.base.rdb seq 2 type b}
{file appendonly.aof.2.incr.aof seq 2 type i}
}
stop_write_load $load_handle0
wait_load_handlers_disconnected
set d1 [r debug digest]
r debug loadaof
set d2 [r debug digest]
assert {$d1 eq $d2}
}
}
Implement Multi Part AOF mechanism to avoid AOFRW overheads. (#9788) Implement Multi-Part AOF mechanism to avoid overheads during AOFRW. Introducing a folder with multiple AOF files tracked by a manifest file. The main issues with the the original AOFRW mechanism are: * buffering of commands that are processed during rewrite (consuming a lot of RAM) * freezes of the main process when the AOFRW completes to drain the remaining part of the buffer and fsync it. * double disk IO for the data that arrives during AOFRW (had to be written to both the old and new AOF files) The main modifications of this PR: 1. Remove the AOF rewrite buffer and related code. 2. Divide the AOF into multiple files, they are classified as two types, one is the the `BASE` type, it represents the full amount of data (Maybe AOF or RDB format) after each AOFRW, there is only one `BASE` file at most. The second is `INCR` type, may have more than one. They represent the incremental commands since the last AOFRW. 3. Use a AOF manifest file to record and manage these AOF files mentioned above. 4. The original configuration of `appendfilename` will be the base part of the new file name, for example: `appendonly.aof.1.base.rdb` and `appendonly.aof.2.incr.aof` 5. Add manifest-related TCL tests, and modified some existing tests that depend on the `appendfilename` 6. Remove the `aof_rewrite_buffer_length` field in info. 7. Add `aof-disable-auto-gc` configuration. By default we're automatically deleting HISTORY type AOFs. It also gives users the opportunity to preserve the history AOFs. just for testing use now. 8. Add AOFRW limiting measure. When the AOFRW failures reaches the threshold (3 times now), we will delay the execution of the next AOFRW by 1 minute. If the next AOFRW also fails, it will be delayed by 2 minutes. The next is 4, 8, 16, the maximum delay is 60 minutes (1 hour). During the limit period, we can still use the 'bgrewriteaof' command to execute AOFRW immediately. 9. Support upgrade (load) data from old version redis. 10. Add `appenddirname` configuration, as the directory name of the append only files. All AOF files and manifest file will be placed in this directory. 11. Only the last AOF file (BASE or INCR) can be truncated. Otherwise redis will exit even if `aof-load-truncated` is enabled. Co-authored-by: Oran Agra <oran@redislabs.com>
2022-01-03 12:14:13 -05:00
}
}