Merge pull request #6247 from oranagra/modules_fork

Module API for Forking

runtest-moduleapi

@@ -13,4 +13,4 @@ then
 fi
 
 make -C tests/modules && \
-$TCLSH tests/test_helper.tcl --single unit/moduleapi/commandfilter --single unit/moduleapi/testrdb "${@}"
+$TCLSH tests/test_helper.tcl --single unit/moduleapi/commandfilter --single unit/moduleapi/fork --single unit/moduleapi/testrdb "${@}"

src/aof.c

@@ -264,9 +264,9 @@ int startAppendOnly(void) {
             strerror(errno));
         return C_ERR;
     }
-    if (server.rdb_child_pid != -1) {
+    if (hasForkChild() && server.aof_child_pid == -1) {
         server.aof_rewrite_scheduled = 1;
-        serverLog(LL_WARNING,"AOF was enabled but there is already a child process saving an RDB file on disk. An AOF background was scheduled to start when possible.");
+        serverLog(LL_WARNING,"AOF was enabled but there is already another background operation. An AOF background was scheduled to start when possible.");
     } else {
         /* If there is a pending AOF rewrite, we need to switch it off and
          * start a new one: the old one cannot be reused because it is not
@@ -395,7 +395,7 @@ void flushAppendOnlyFile(int force) {
      * useful for graphing / monitoring purposes. */
     if (sync_in_progress) {
         latencyAddSampleIfNeeded("aof-write-pending-fsync",latency);
-    } else if (server.aof_child_pid != -1 || server.rdb_child_pid != -1) {
+    } else if (hasForkChild()) {
         latencyAddSampleIfNeeded("aof-write-active-child",latency);
     } else {
         latencyAddSampleIfNeeded("aof-write-alone",latency);
@@ -491,9 +491,8 @@ void flushAppendOnlyFile(int force) {
 try_fsync:
     /* Don't fsync if no-appendfsync-on-rewrite is set to yes and there are
      * children doing I/O in the background. */
-    if (server.aof_no_fsync_on_rewrite &&
+    if (server.aof_no_fsync_on_rewrite && hasForkChild())
-        (server.aof_child_pid != -1 || server.rdb_child_pid != -1))
+        return;
-            return;
 
     /* Perform the fsync if needed. */
     if (server.aof_fsync == AOF_FSYNC_ALWAYS) {
@@ -1563,39 +1562,24 @@ void aofClosePipes(void) {
  */
 int rewriteAppendOnlyFileBackground(void) {
     pid_t childpid;
-    long long start;
 
-    if (server.aof_child_pid != -1 || server.rdb_child_pid != -1) return C_ERR;
+    if (hasForkChild()) return C_ERR;
     if (aofCreatePipes() != C_OK) return C_ERR;
     openChildInfoPipe();
-    start = ustime();
+    if ((childpid = redisFork()) == 0) {
-    if ((childpid = fork()) == 0) {
         char tmpfile[256];
 
         /* Child */
-        closeListeningSockets(0);
         redisSetProcTitle("redis-aof-rewrite");
         snprintf(tmpfile,256,"temp-rewriteaof-bg-%d.aof", (int) getpid());
         if (rewriteAppendOnlyFile(tmpfile) == C_OK) {
-            size_t private_dirty = zmalloc_get_private_dirty(-1);
+            sendChildCOWInfo(CHILD_INFO_TYPE_AOF, "AOF rewrite");
-
-            if (private_dirty) {
-                serverLog(LL_NOTICE,
-                    "AOF rewrite: %zu MB of memory used by copy-on-write",
-                    private_dirty/(1024*1024));
-            }
-
-            server.child_info_data.cow_size = private_dirty;
-            sendChildInfo(CHILD_INFO_TYPE_AOF);
             exitFromChild(0);
         } else {
             exitFromChild(1);
         }
     } else {
         /* Parent */
-        server.stat_fork_time = ustime()-start;
-        server.stat_fork_rate = (double) zmalloc_used_memory() * 1000000 / server.stat_fork_time / (1024*1024*1024); /* GB per second. */
-        latencyAddSampleIfNeeded("fork",server.stat_fork_time/1000);
         if (childpid == -1) {
             closeChildInfoPipe();
             serverLog(LL_WARNING,
@@ -1609,7 +1593,6 @@ int rewriteAppendOnlyFileBackground(void) {
         server.aof_rewrite_scheduled = 0;
         server.aof_rewrite_time_start = time(NULL);
         server.aof_child_pid = childpid;
-        updateDictResizePolicy();
         /* We set appendseldb to -1 in order to force the next call to the
          * feedAppendOnlyFile() to issue a SELECT command, so the differences
          * accumulated by the parent into server.aof_rewrite_buf will start
@@ -1624,7 +1607,7 @@ int rewriteAppendOnlyFileBackground(void) {
 void bgrewriteaofCommand(client *c) {
     if (server.aof_child_pid != -1) {
         addReplyError(c,"Background append only file rewriting already in progress");
-    } else if (server.rdb_child_pid != -1) {
+    } else if (hasForkChild()) {
         server.aof_rewrite_scheduled = 1;
         addReplyStatus(c,"Background append only file rewriting scheduled");
     } else if (rewriteAppendOnlyFileBackground() == C_OK) {

src/childinfo.c

@@ -80,6 +80,8 @@ void receiveChildInfo(void) {
             server.stat_rdb_cow_bytes = server.child_info_data.cow_size;
         } else if (server.child_info_data.process_type == CHILD_INFO_TYPE_AOF) {
             server.stat_aof_cow_bytes = server.child_info_data.cow_size;
+        } else if (server.child_info_data.process_type == CHILD_INFO_TYPE_MODULE) {
+            server.stat_module_cow_bytes = server.child_info_data.cow_size;
         }
     }
 }

src/db.c

@@ -60,10 +60,7 @@ robj *lookupKey(redisDb *db, robj *key, int flags) {
         /* Update the access time for the ageing algorithm.
          * Don't do it if we have a saving child, as this will trigger
          * a copy on write madness. */
-        if (server.rdb_child_pid == -1 &&
+        if (!hasForkChild() && !(flags & LOOKUP_NOTOUCH)){
-            server.aof_child_pid == -1 &&
-            !(flags & LOOKUP_NOTOUCH))
-        {
             if (server.maxmemory_policy & MAXMEMORY_FLAG_LFU) {
                 updateLFU(val);
             } else {

src/defrag.c

@@ -1039,7 +1039,7 @@ void activeDefragCycle(void) {
     mstime_t latency;
     int quit = 0;
 
-    if (server.aof_child_pid!=-1 || server.rdb_child_pid!=-1)
+    if (hasForkChild())
         return; /* Defragging memory while there's a fork will just do damage. */
 
     /* Once a second, check if we the fragmentation justfies starting a scan

src/module.c

@@ -31,6 +31,7 @@
 #include "cluster.h"
 #include "rdb.h"
 #include <dlfcn.h>
+#include <wait.h>
 
 #define REDISMODULE_CORE 1
 #include "redismodule.h"
@@ -293,6 +294,14 @@ typedef struct RedisModuleCommandFilter {
 /* Registered filters */
 static list *moduleCommandFilters;
 
+typedef void (*RedisModuleForkDoneHandler) (int exitcode, int bysignal, void *user_data);
+
+static struct RedisModuleForkInfo {
+    RedisModuleForkDoneHandler done_handler;
+    void* done_handler_user_data;
+} moduleForkInfo = {0};
+
+
 /* --------------------------------------------------------------------------
  * Prototypes
  * -------------------------------------------------------------------------- */
@@ -5130,6 +5139,98 @@ int RM_CommandFilterArgDelete(RedisModuleCommandFilterCtx *fctx, int pos)
     return REDISMODULE_OK;
 }
 
+/* --------------------------------------------------------------------------
+ * Module fork API
+ * -------------------------------------------------------------------------- */
+
+/* Create a background child process with the current frozen snaphost of the
+ * main process where you can do some processing in the background without
+ * affecting / freezing the traffic and no need for threads and GIL locking.
+ * Note that Redis allows for only one concurrent fork.
+ * When the child wants to exit, it should call RedisModule_ExitFromChild.
+ * If the parent wants to kill the child it should call RedisModule_KillForkChild
+ * The done handler callback will be executed on the parent process when the
+ * child existed (but not when killed)
+ * Return: -1 on failure, on success the parent process will get a positive PID
+ * of the child, and the child process will get 0.
+ */
+int RM_Fork(RedisModuleForkDoneHandler cb, void *user_data)
+{
+    pid_t childpid;
+    if (hasForkChild()) {
+        return -1;
+    }
+
+    openChildInfoPipe();
+    if ((childpid = redisFork()) == 0) {
+        /* Child */
+        redisSetProcTitle("redis-module-fork");
+    } else if (childpid == -1) {
+        closeChildInfoPipe();
+        serverLog(LL_WARNING,"Can't fork for module: %s", strerror(errno));
+    } else {
+        /* Parent */
+        server.module_child_pid = childpid;
+        moduleForkInfo.done_handler = cb;
+        moduleForkInfo.done_handler_user_data = user_data;
+        serverLog(LL_NOTICE, "Module fork started pid: %d ", childpid);
+    }
+    return childpid;
+}
+
+/* Call from the child process when you want to terminate it.
+ * retcode will be provided to the done handler executed on the parent process.
+ */
+int RM_ExitFromChild(int retcode)
+{
+    sendChildCOWInfo(CHILD_INFO_TYPE_MODULE, "Module fork");
+    exitFromChild(retcode);
+    return REDISMODULE_OK;
+}
+
+void TerminateModuleForkChild(int wait) {
+    int statloc;
+    serverLog(LL_NOTICE,"Killing running module fork child: %ld",
+        (long) server.module_child_pid);
+    if (kill(server.module_child_pid,SIGUSR1) != -1 && wait) {
+        while(wait3(&statloc,0,NULL) != server.module_child_pid);
+    }
+    /* Reset the buffer accumulating changes while the child saves. */
+    server.module_child_pid = -1;
+    moduleForkInfo.done_handler = NULL;
+    moduleForkInfo.done_handler_user_data = NULL;
+    closeChildInfoPipe();
+    updateDictResizePolicy();
+}
+
+/* Can be used to kill the forked child process from the parent process.
+ * child_pid whould be the return value of RedisModule_Fork. */
+int RM_KillForkChild(int child_pid)
+{
+    /* No module child? return. */
+    if (server.module_child_pid == -1) return REDISMODULE_ERR;
+    /* Make sure the module knows the pid it wants to kill (not trying to
+     * randomly kill other module's forks) */
+    if (server.module_child_pid != child_pid) return REDISMODULE_ERR;
+    /* Kill module child, wait for child exit. */
+    TerminateModuleForkChild(1);
+    return REDISMODULE_OK;
+}
+
+void ModuleForkDoneHandler(int exitcode, int bysignal)
+{
+    serverLog(LL_NOTICE,
+        "Module fork exited pid: %d, retcode: %d, bysignal: %d",
+        server.module_child_pid, exitcode, bysignal);
+    if (moduleForkInfo.done_handler) {
+        moduleForkInfo.done_handler(exitcode, bysignal,
+            moduleForkInfo.done_handler_user_data);
+    }
+    server.module_child_pid = -1;
+    moduleForkInfo.done_handler = NULL;
+    moduleForkInfo.done_handler_user_data = NULL;
+}
+
 /* --------------------------------------------------------------------------
  * Modules API internals
  * -------------------------------------------------------------------------- */
@@ -5652,4 +5753,7 @@ void moduleRegisterCoreAPI(void) {
     REGISTER_API(CommandFilterArgInsert);
     REGISTER_API(CommandFilterArgReplace);
     REGISTER_API(CommandFilterArgDelete);
+    REGISTER_API(Fork);
+    REGISTER_API(ExitFromChild);
+    REGISTER_API(KillForkChild);
 }

src/rdb.c

@@ -1335,40 +1335,25 @@ werr:
 
 int rdbSaveBackground(char *filename, rdbSaveInfo *rsi) {
     pid_t childpid;
-    long long start;
 
-    if (server.aof_child_pid != -1 || server.rdb_child_pid != -1) return C_ERR;
+    if (hasForkChild()) return C_ERR;
 
     server.dirty_before_bgsave = server.dirty;
     server.lastbgsave_try = time(NULL);
     openChildInfoPipe();
 
-    start = ustime();
+    if ((childpid = redisFork()) == 0) {
-    if ((childpid = fork()) == 0) {
         int retval;
 
         /* Child */
-        closeListeningSockets(0);
         redisSetProcTitle("redis-rdb-bgsave");
         retval = rdbSave(filename,rsi);
         if (retval == C_OK) {
-            size_t private_dirty = zmalloc_get_private_dirty(-1);
+            sendChildCOWInfo(CHILD_INFO_TYPE_RDB, "RDB");
-
-            if (private_dirty) {
-                serverLog(LL_NOTICE,
-                    "RDB: %zu MB of memory used by copy-on-write",
-                    private_dirty/(1024*1024));
-            }
-
-            server.child_info_data.cow_size = private_dirty;
-            sendChildInfo(CHILD_INFO_TYPE_RDB);
         }
         exitFromChild((retval == C_OK) ? 0 : 1);
     } else {
         /* Parent */
-        server.stat_fork_time = ustime()-start;
-        server.stat_fork_rate = (double) zmalloc_used_memory() * 1000000 / server.stat_fork_time / (1024*1024*1024); /* GB per second. */
-        latencyAddSampleIfNeeded("fork",server.stat_fork_time/1000);
         if (childpid == -1) {
             closeChildInfoPipe();
             server.lastbgsave_status = C_ERR;
@@ -1380,7 +1365,6 @@ int rdbSaveBackground(char *filename, rdbSaveInfo *rsi) {
         server.rdb_save_time_start = time(NULL);
         server.rdb_child_pid = childpid;
         server.rdb_child_type = RDB_CHILD_TYPE_DISK;
-        updateDictResizePolicy();
         return C_OK;
     }
     return C_OK; /* unreached */
@@ -2431,10 +2415,9 @@ int rdbSaveToSlavesSockets(rdbSaveInfo *rsi) {
     listNode *ln;
     listIter li;
     pid_t childpid;
-    long long start;
     int pipefds[2];
 
-    if (server.aof_child_pid != -1 || server.rdb_child_pid != -1) return C_ERR;
+    if (hasForkChild()) return C_ERR;
 
     /* Before to fork, create a pipe that will be used in order to
      * send back to the parent the IDs of the slaves that successfully
@@ -2470,8 +2453,7 @@ int rdbSaveToSlavesSockets(rdbSaveInfo *rsi) {
 
     /* Create the child process. */
     openChildInfoPipe();
-    start = ustime();
+    if ((childpid = redisFork()) == 0) {
-    if ((childpid = fork()) == 0) {
         /* Child */
         int retval;
         rio slave_sockets;
@@ -2479,7 +2461,6 @@ int rdbSaveToSlavesSockets(rdbSaveInfo *rsi) {
         rioInitWithFdset(&slave_sockets,fds,numfds);
         zfree(fds);
 
-        closeListeningSockets(0);
         redisSetProcTitle("redis-rdb-to-slaves");
 
         retval = rdbSaveRioWithEOFMark(&slave_sockets,NULL,rsi);
@@ -2487,16 +2468,7 @@ int rdbSaveToSlavesSockets(rdbSaveInfo *rsi) {
             retval = C_ERR;
 
         if (retval == C_OK) {
-            size_t private_dirty = zmalloc_get_private_dirty(-1);
+            sendChildCOWInfo(CHILD_INFO_TYPE_RDB, "RDB");
-
-            if (private_dirty) {
-                serverLog(LL_NOTICE,
-                    "RDB: %zu MB of memory used by copy-on-write",
-                    private_dirty/(1024*1024));
-            }
-
-            server.child_info_data.cow_size = private_dirty;
-            sendChildInfo(CHILD_INFO_TYPE_RDB);
 
             /* If we are returning OK, at least one slave was served
              * with the RDB file as expected, so we need to send a report
@@ -2565,16 +2537,11 @@ int rdbSaveToSlavesSockets(rdbSaveInfo *rsi) {
             close(pipefds[1]);
             closeChildInfoPipe();
         } else {
-            server.stat_fork_time = ustime()-start;
-            server.stat_fork_rate = (double) zmalloc_used_memory() * 1000000 / server.stat_fork_time / (1024*1024*1024); /* GB per second. */
-            latencyAddSampleIfNeeded("fork",server.stat_fork_time/1000);
-
             serverLog(LL_NOTICE,"Background RDB transfer started by pid %d",
                 childpid);
             server.rdb_save_time_start = time(NULL);
             server.rdb_child_pid = childpid;
             server.rdb_child_type = RDB_CHILD_TYPE_SOCKET;
-            updateDictResizePolicy();
         }
         zfree(clientids);
         zfree(fds);
@@ -2617,13 +2584,13 @@ void bgsaveCommand(client *c) {
 
     if (server.rdb_child_pid != -1) {
         addReplyError(c,"Background save already in progress");
-    } else if (server.aof_child_pid != -1) {
+    } else if (hasForkChild()) {
         if (schedule) {
             server.rdb_bgsave_scheduled = 1;
             addReplyStatus(c,"Background saving scheduled");
         } else {
             addReplyError(c,
-                "An AOF log rewriting in progress: can't BGSAVE right now. "
+                "Another BG operation is in progress: can't BGSAVE right now. "
                 "Use BGSAVE SCHEDULE in order to schedule a BGSAVE whenever "
                 "possible.");
         }

src/redismodule.h

@@ -182,6 +182,7 @@ typedef void (*RedisModuleTypeFreeFunc)(void *value);
 typedef void (*RedisModuleClusterMessageReceiver)(RedisModuleCtx *ctx, const char *sender_id, uint8_t type, const unsigned char *payload, uint32_t len);
 typedef void (*RedisModuleTimerProc)(RedisModuleCtx *ctx, void *data);
 typedef void (*RedisModuleCommandFilterFunc) (RedisModuleCommandFilterCtx *filter);
+typedef void (*RedisModuleForkDoneHandler) (int exitcode, int bysignal, void *user_data);
 
 #define REDISMODULE_TYPE_METHOD_VERSION 2
 typedef struct RedisModuleTypeMethods {
@@ -372,6 +373,9 @@ const RedisModuleString *REDISMODULE_API_FUNC(RedisModule_CommandFilterArgGet)(R
 int REDISMODULE_API_FUNC(RedisModule_CommandFilterArgInsert)(RedisModuleCommandFilterCtx *fctx, int pos, RedisModuleString *arg);
 int REDISMODULE_API_FUNC(RedisModule_CommandFilterArgReplace)(RedisModuleCommandFilterCtx *fctx, int pos, RedisModuleString *arg);
 int REDISMODULE_API_FUNC(RedisModule_CommandFilterArgDelete)(RedisModuleCommandFilterCtx *fctx, int pos);
+int REDISMODULE_API_FUNC(RedisModule_Fork)(RedisModuleForkDoneHandler cb, void *user_data);
+int REDISMODULE_API_FUNC(RedisModule_ExitFromChild)(int retcode);
+int REDISMODULE_API_FUNC(RedisModule_KillForkChild)(int child_pid);
 #endif
 
 /* This is included inline inside each Redis module. */
@@ -546,6 +550,9 @@ static int RedisModule_Init(RedisModuleCtx *ctx, const char *name, int ver, int
     REDISMODULE_GET_API(CommandFilterArgInsert);
     REDISMODULE_GET_API(CommandFilterArgReplace);
     REDISMODULE_GET_API(CommandFilterArgDelete);
+    REDISMODULE_GET_API(Fork);
+    REDISMODULE_GET_API(ExitFromChild);
+    REDISMODULE_GET_API(KillForkChild);
 #endif
 
     if (RedisModule_IsModuleNameBusy && RedisModule_IsModuleNameBusy(name)) return REDISMODULE_ERR;

src/replication.c

@@ -751,11 +751,11 @@ void syncCommand(client *c) {
             /* Target is disk (or the slave is not capable of supporting
              * diskless replication) and we don't have a BGSAVE in progress,
              * let's start one. */
-            if (server.aof_child_pid == -1) {
+            if (!hasForkChild()) {
                 startBgsaveForReplication(c->slave_capa);
             } else {
                 serverLog(LL_NOTICE,
-                    "No BGSAVE in progress, but an AOF rewrite is active. "
+                    "No BGSAVE in progress, but another BG operation is active. "
                     "BGSAVE for replication delayed");
             }
         }
@@ -2930,7 +2930,7 @@ void replicationCron(void) {
      * In case of diskless replication, we make sure to wait the specified
      * number of seconds (according to configuration) so that other slaves
      * have the time to arrive before we start streaming. */
-    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
+    if (!hasForkChild()) {
         time_t idle, max_idle = 0;
         int slaves_waiting = 0;
         int mincapa = -1;

src/scripting.c

@@ -1827,7 +1827,7 @@ void ldbSendLogs(void) {
 int ldbStartSession(client *c) {
     ldb.forked = (c->flags & CLIENT_LUA_DEBUG_SYNC) == 0;
     if (ldb.forked) {
-        pid_t cp = fork();
+        pid_t cp = redisFork();
         if (cp == -1) {
             addReplyError(c,"Fork() failed: can't run EVAL in debugging mode.");
             return 0;
@@ -1844,7 +1844,6 @@ int ldbStartSession(client *c) {
              * socket to make sure if the parent crashes a reset is sent
              * to the clients. */
             serverLog(LL_WARNING,"Redis forked for debugging eval");
-            closeListeningSockets(0);
         } else {
             /* Parent */
             listAddNodeTail(ldb.children,(void*)(unsigned long)cp);

src/server.c

@@ -1449,12 +1449,18 @@ int incrementallyRehash(int dbid) {
  * for dict.c to resize the hash tables accordingly to the fact we have o not
  * running childs. */
 void updateDictResizePolicy(void) {
-    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1)
+    if (!hasForkChild())
         dictEnableResize();
     else
         dictDisableResize();
 }
 
+int hasForkChild() {
+    return server.rdb_child_pid != -1 ||
+           server.aof_child_pid != -1 ||
+           server.module_child_pid != -1;
+}
+
 /* ======================= Cron: called every 100 ms ======================== */
 
 /* Add a sample to the operations per second array of samples. */
@@ -1691,7 +1697,7 @@ void databasesCron(void) {
     /* Perform hash tables rehashing if needed, but only if there are no
      * other processes saving the DB on disk. Otherwise rehashing is bad
      * as will cause a lot of copy-on-write of memory pages. */
-    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1) {
+    if (!hasForkChild()) {
         /* We use global counters so if we stop the computation at a given
          * DB we'll be able to start from the successive in the next
          * cron loop iteration. */
@@ -1888,15 +1894,14 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
 
     /* Start a scheduled AOF rewrite if this was requested by the user while
      * a BGSAVE was in progress. */
-    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
+    if (!hasForkChild() &&
         server.aof_rewrite_scheduled)
     {
         rewriteAppendOnlyFileBackground();
     }
 
     /* Check if a background saving or AOF rewrite in progress terminated. */
-    if (server.rdb_child_pid != -1 || server.aof_child_pid != -1 ||
+    if (hasForkChild() || ldbPendingChildren())
-        ldbPendingChildren())
     {
         int statloc;
         pid_t pid;
@@ -1907,18 +1912,29 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
 
             if (WIFSIGNALED(statloc)) bysignal = WTERMSIG(statloc);
 
+            /* sigKillChildHandler catches the signal and calls exit(), but we
+             * must make sure not to flag lastbgsave_status, etc incorrectly. */
+            if (exitcode == SIGUSR1) {
+                bysignal = SIGUSR1;
+                exitcode = 1;
+            }
+
             if (pid == -1) {
                 serverLog(LL_WARNING,"wait3() returned an error: %s. "
-                    "rdb_child_pid = %d, aof_child_pid = %d",
+                    "rdb_child_pid = %d, aof_child_pid = %d, module_child_pid = %d",
                     strerror(errno),
                     (int) server.rdb_child_pid,
-                    (int) server.aof_child_pid);
+                    (int) server.aof_child_pid,
+                    (int) server.module_child_pid);
             } else if (pid == server.rdb_child_pid) {
                 backgroundSaveDoneHandler(exitcode,bysignal);
                 if (!bysignal && exitcode == 0) receiveChildInfo();
             } else if (pid == server.aof_child_pid) {
                 backgroundRewriteDoneHandler(exitcode,bysignal);
                 if (!bysignal && exitcode == 0) receiveChildInfo();
+            } else if (pid == server.module_child_pid) {
+                ModuleForkDoneHandler(exitcode,bysignal);
+                if (!bysignal && exitcode == 0) receiveChildInfo();
             } else {
                 if (!ldbRemoveChild(pid)) {
                     serverLog(LL_WARNING,
@@ -1956,8 +1972,7 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
 
         /* Trigger an AOF rewrite if needed. */
         if (server.aof_state == AOF_ON &&
-            server.rdb_child_pid == -1 &&
+            !hasForkChild() &&
-            server.aof_child_pid == -1 &&
             server.aof_rewrite_perc &&
             server.aof_current_size > server.aof_rewrite_min_size)
         {
@@ -2015,7 +2030,7 @@ int serverCron(struct aeEventLoop *eventLoop, long long id, void *clientData) {
      * Note: this code must be after the replicationCron() call above so
      * make sure when refactoring this file to keep this order. This is useful
      * because we want to give priority to RDB savings for replication. */
-    if (server.rdb_child_pid == -1 && server.aof_child_pid == -1 &&
+    if (!hasForkChild() &&
         server.rdb_bgsave_scheduled &&
         (server.unixtime-server.lastbgsave_try > CONFIG_BGSAVE_RETRY_DELAY ||
          server.lastbgsave_status == C_OK))
@@ -2799,6 +2814,7 @@ void initServer(void) {
     server.cronloops = 0;
     server.rdb_child_pid = -1;
     server.aof_child_pid = -1;
+    server.module_child_pid = -1;
     server.rdb_child_type = RDB_CHILD_TYPE_NONE;
     server.rdb_bgsave_scheduled = 0;
     server.child_info_pipe[0] = -1;
@@ -2817,6 +2833,7 @@ void initServer(void) {
     server.stat_peak_memory = 0;
     server.stat_rdb_cow_bytes = 0;
     server.stat_aof_cow_bytes = 0;
+    server.stat_module_cow_bytes = 0;
     server.cron_malloc_stats.zmalloc_used = 0;
     server.cron_malloc_stats.process_rss = 0;
     server.cron_malloc_stats.allocator_allocated = 0;
@@ -3566,6 +3583,12 @@ int prepareForShutdown(int flags) {
         killRDBChild();
     }
 
+    /* Kill module child if there is one. */
+    if (server.module_child_pid != -1) {
+        serverLog(LL_WARNING,"There is a module fork child. Killing it!");
+        TerminateModuleForkChild(0);
+    }
+
     if (server.aof_state != AOF_OFF) {
         /* Kill the AOF saving child as the AOF we already have may be longer
          * but contains the full dataset anyway. */
@@ -4066,7 +4089,9 @@ sds genRedisInfoString(char *section) {
             "aof_current_rewrite_time_sec:%jd\r\n"
             "aof_last_bgrewrite_status:%s\r\n"
             "aof_last_write_status:%s\r\n"
-            "aof_last_cow_size:%zu\r\n",
+            "aof_last_cow_size:%zu\r\n"
+            "module_fork_in_progress:%d\r\n"
+            "module_fork_last_cow_size:%zu\r\n",
             server.loading,
             server.dirty,
             server.rdb_child_pid != -1,
@@ -4084,7 +4109,9 @@ sds genRedisInfoString(char *section) {
                 -1 : time(NULL)-server.aof_rewrite_time_start),
             (server.aof_lastbgrewrite_status == C_OK) ? "ok" : "err",
             (server.aof_last_write_status == C_OK) ? "ok" : "err",
-            server.stat_aof_cow_bytes);
+            server.stat_aof_cow_bytes,
+            server.module_child_pid != -1,
+            server.stat_module_cow_bytes);
 
         if (server.aof_enabled) {
             info = sdscatprintf(info,
@@ -4591,6 +4618,58 @@ void setupSignalHandlers(void) {
     return;
 }
 
+static void sigKillChildHandler(int sig) {
+    UNUSED(sig);
+    /* this handler is needed to resolve a valgrind warning */
+    serverLogFromHandler(LL_WARNING, "Received SIGUSR1 in child, exiting now.");
+    exitFromChild(SIGUSR1);
+}
+
+void setupChildSignalHandlers(void) {
+    struct sigaction act;
+
+    /* When the SA_SIGINFO flag is set in sa_flags then sa_sigaction is used.
+     * Otherwise, sa_handler is used. */
+    sigemptyset(&act.sa_mask);
+    act.sa_flags = 0;
+    act.sa_handler = sigKillChildHandler;
+    sigaction(SIGUSR1, &act, NULL);
+    return;
+}
+
+int redisFork() {
+    int childpid;
+    long long start = ustime();
+    if ((childpid = fork()) == 0) {
+        /* Child */
+        closeListeningSockets(0);
+        setupChildSignalHandlers();
+    } else {
+        /* Parent */
+        server.stat_fork_time = ustime()-start;
+        server.stat_fork_rate = (double) zmalloc_used_memory() * 1000000 / server.stat_fork_time / (1024*1024*1024); /* GB per second. */
+        latencyAddSampleIfNeeded("fork",server.stat_fork_time/1000);
+        if (childpid == -1) {
+            return -1;
+        }
+        updateDictResizePolicy();
+    }
+    return childpid;
+}
+
+void sendChildCOWInfo(int ptype, char *pname) {
+    size_t private_dirty = zmalloc_get_private_dirty(-1);
+
+    if (private_dirty) {
+        serverLog(LL_NOTICE,
+            "%s: %zu MB of memory used by copy-on-write",
+            pname, private_dirty/(1024*1024));
+    }
+
+    server.child_info_data.cow_size = private_dirty;
+    sendChildInfo(ptype);
+}
+
 void memtest(size_t megabytes, int passes);
 
 /* Returns 1 if there is --sentinel among the arguments or if

src/server.h

@@ -1041,6 +1041,7 @@ struct clusterState;
 #define CHILD_INFO_MAGIC 0xC17DDA7A12345678LL
 #define CHILD_INFO_TYPE_RDB 0
 #define CHILD_INFO_TYPE_AOF 1
+#define CHILD_INFO_TYPE_MODULE 3
 
 struct redisServer {
     /* General */
@@ -1076,6 +1077,7 @@ struct redisServer {
     int module_blocked_pipe[2]; /* Pipe used to awake the event loop if a
                                    client blocked on a module command needs
                                    to be processed. */
+    pid_t module_child_pid;     /* PID of module child */
     /* Networking */
     int port;                   /* TCP listening port */
     int tcp_backlog;            /* TCP listen() backlog */
@@ -1149,6 +1151,7 @@ struct redisServer {
     _Atomic long long stat_net_output_bytes; /* Bytes written to network. */
     size_t stat_rdb_cow_bytes;      /* Copy on write bytes during RDB saving. */
     size_t stat_aof_cow_bytes;      /* Copy on write bytes during AOF rewrite. */
+    size_t stat_module_cow_bytes;   /* Copy on write bytes during module fork. */
     /* The following two are used to track instantaneous metrics, like
      * number of operations per second, network traffic. */
     struct {
@@ -1540,6 +1543,8 @@ void moduleAcquireGIL(void);
 void moduleReleaseGIL(void);
 void moduleNotifyKeyspaceEvent(int type, const char *event, robj *key, int dbid);
 void moduleCallCommandFilters(client *c);
+void ModuleForkDoneHandler(int exitcode, int bysignal);
+void TerminateModuleForkChild(int wait);
 ssize_t rdbSaveModulesAux(rio *rdb, int when);
 int moduleAllDatatypesHandleErrors();
 
@@ -1803,6 +1808,11 @@ void closeChildInfoPipe(void);
 void sendChildInfo(int process_type);
 void receiveChildInfo(void);
 
+/* Fork helpers */
+int redisFork();
+int hasForkChild();
+void sendChildCOWInfo(int ptype, char *pname);
+
 /* acl.c -- Authentication related prototypes. */
 extern rax *Users;
 extern user *DefaultUser;

tests/integration/rdb.tcl

@@ -115,3 +115,17 @@ start_server_and_kill_it [list "dir" $server_path] {
         }
     }
 }
+
+start_server {} {
+    test {Test FLUSHALL aborts bgsave} {
+        r config set rdb-key-save-delay 1000
+        r debug populate 1000
+        r bgsave
+        assert_equal [s rdb_bgsave_in_progress] 1
+        r flushall
+        after 200
+        assert_equal [s rdb_bgsave_in_progress] 0
+        # make sure the server is still writable
+        r set x xx
+    }
+}

tests/modules/Makefile

@@ -13,16 +13,20 @@ endif
 
 .SUFFIXES: .c .so .xo .o
 
-all: commandfilter.so testrdb.so
+all: commandfilter.so testrdb.so fork.so
 
 .c.xo:
 	$(CC) -I../../src $(CFLAGS) $(SHOBJ_CFLAGS) -fPIC -c $< -o $@
 
 commandfilter.xo: ../../src/redismodule.h
+fork.xo: ../../src/redismodule.h
 testrdb.xo: ../../src/redismodule.h
 
 commandfilter.so: commandfilter.xo
 	$(LD) -o $@ $< $(SHOBJ_LDFLAGS) $(LIBS) -lc
 
+fork.so: fork.xo
+	$(LD) -o $@ $< $(SHOBJ_LDFLAGS) $(LIBS) -lc
+
 testrdb.so: testrdb.xo
 	$(LD) -o $@ $< $(SHOBJ_LDFLAGS) $(LIBS) -lc

tests/modules/fork.c

@@ -0,0 +1,84 @@
+#define REDISMODULE_EXPERIMENTAL_API
+#include "redismodule.h"
+
+#include <string.h>
+#include <assert.h>
+#include <unistd.h>
+
+#define UNUSED(V) ((void) V)
+
+int child_pid = -1;
+int exitted_with_code = -1;
+
+void done_handler(int exitcode, int bysignal, void *user_data) {
+    child_pid = -1;
+    exitted_with_code = exitcode;
+    assert(user_data==(void*)0xdeadbeef);
+    UNUSED(bysignal);
+}
+
+int fork_create(RedisModuleCtx *ctx, RedisModuleString **argv, int argc)
+{
+    long long code_to_exit_with;
+    if (argc != 2) {
+        RedisModule_WrongArity(ctx);
+        return REDISMODULE_OK;
+    }
+    RedisModule_StringToLongLong(argv[1], &code_to_exit_with);
+    exitted_with_code = -1;
+    child_pid = RedisModule_Fork(done_handler, (void*)0xdeadbeef);
+    if (child_pid < 0) {
+        RedisModule_ReplyWithError(ctx, "Fork failed");
+        return REDISMODULE_OK;
+    } else if (child_pid > 0) {
+        /* parent */
+        RedisModule_ReplyWithLongLong(ctx, child_pid);
+        return REDISMODULE_OK;
+    }
+
+    /* child */
+    RedisModule_Log(ctx, "notice", "fork child started");
+    usleep(200000);
+    RedisModule_Log(ctx, "notice", "fork child exiting");
+    RedisModule_ExitFromChild(code_to_exit_with);
+    /* unreachable */
+    return 0;
+}
+
+int fork_exitcode(RedisModuleCtx *ctx, RedisModuleString **argv, int argc)
+{
+    UNUSED(argv);
+    UNUSED(argc);
+    RedisModule_ReplyWithLongLong(ctx, exitted_with_code);
+    return REDISMODULE_OK;
+}
+
+int fork_kill(RedisModuleCtx *ctx, RedisModuleString **argv, int argc)
+{
+    UNUSED(argv);
+    UNUSED(argc);
+    if (RedisModule_KillForkChild(child_pid) != REDISMODULE_OK)
+        RedisModule_ReplyWithError(ctx, "KillForkChild failed");
+    else
+        RedisModule_ReplyWithLongLong(ctx, 1);
+    child_pid = -1;
+    return REDISMODULE_OK;
+}
+
+int RedisModule_OnLoad(RedisModuleCtx *ctx, RedisModuleString **argv, int argc) {
+    UNUSED(argv);
+    UNUSED(argc);
+    if (RedisModule_Init(ctx,"fork",1,REDISMODULE_APIVER_1)== REDISMODULE_ERR)
+        return REDISMODULE_ERR;
+
+    if (RedisModule_CreateCommand(ctx,"fork.create", fork_create,"",0,0,0) == REDISMODULE_ERR)
+        return REDISMODULE_ERR;
+
+    if (RedisModule_CreateCommand(ctx,"fork.exitcode", fork_exitcode,"",0,0,0) == REDISMODULE_ERR)
+        return REDISMODULE_ERR;
+
+    if (RedisModule_CreateCommand(ctx,"fork.kill", fork_kill,"",0,0,0) == REDISMODULE_ERR)
+        return REDISMODULE_ERR;
+
+    return REDISMODULE_OK;
+}

tests/unit/moduleapi/fork.tcl

@@ -0,0 +1,32 @@
+set testmodule [file normalize tests/modules/fork.so]
+
+proc count_log_message {pattern} {
+    set result [exec grep -c $pattern < [srv 0 stdout]]
+}
+
+start_server {tags {"modules"}} {
+    r module load $testmodule
+
+    test {Module fork} {
+        # the argument to fork.create is the exitcode on termination
+        r fork.create 3
+        wait_for_condition 20 100 {
+            [r fork.exitcode] != -1
+        } else {
+            fail "fork didn't terminate"
+        }
+        r fork.exitcode
+    } {3}
+
+    test {Module fork kill} {
+        r fork.create 3
+        after 20
+        r fork.kill
+        after 100
+
+        assert {[count_log_message "fork child started"] eq "2"}
+        assert {[count_log_message "Received SIGUSR1 in child"] eq "1"}
+        assert {[count_log_message "fork child exiting"] eq "1"}
+    }
+
+}