Commit bf759cc9 authored by Oran Agra's avatar Oran Agra
Browse files

Merge remote-tracking branch 'antirez/unstable' into jemalloc_purge_bg

parents 2e19b941 ee1cef18
......@@ -264,9 +264,9 @@ int startAppendOnly(void) {
strerror(errno));
return C_ERR;
}
if (server.rdb_child_pid != -1) {
if (hasActiveChildProcess() && 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
......@@ -303,9 +303,7 @@ ssize_t aofWrite(int fd, const char *buf, size_t len) {
nwritten = write(fd, buf, len);
if (nwritten < 0) {
if (errno == EINTR) {
continue;
}
if (errno == EINTR) continue;
return totwritten ? totwritten : -1;
}
......@@ -397,7 +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 (hasActiveChildProcess()) {
latencyAddSampleIfNeeded("aof-write-active-child",latency);
} else {
latencyAddSampleIfNeeded("aof-write-alone",latency);
......@@ -493,8 +491,7 @@ 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 &&
(server.aof_child_pid != -1 || server.rdb_child_pid != -1))
if (server.aof_no_fsync_on_rewrite && hasActiveChildProcess())
return;
/* Perform the fsync if needed. */
......@@ -729,7 +726,7 @@ int loadAppendOnlyFile(char *filename) {
server.aof_state = AOF_OFF;
fakeClient = createFakeClient();
startLoading(fp);
startLoadingFile(fp, filename);
/* Check if this AOF file has an RDB preamble. In that case we need to
* load the RDB file and later continue loading the AOF tail. */
......@@ -864,6 +861,7 @@ loaded_ok: /* DB loaded, cleanup and return C_OK to the caller. */
readerr: /* Read error. If feof(fp) is true, fall through to unexpected EOF. */
if (!feof(fp)) {
if (fakeClient) freeFakeClient(fakeClient); /* avoid valgrind warning */
fclose(fp);
serverLog(LL_WARNING,"Unrecoverable error reading the append only file: %s", strerror(errno));
exit(1);
}
......@@ -894,11 +892,13 @@ uxeof: /* Unexpected AOF end of file. */
}
}
if (fakeClient) freeFakeClient(fakeClient); /* avoid valgrind warning */
fclose(fp);
serverLog(LL_WARNING,"Unexpected end of file reading the append only file. You can: 1) Make a backup of your AOF file, then use ./redis-check-aof --fix <filename>. 2) Alternatively you can set the 'aof-load-truncated' configuration option to yes and restart the server.");
exit(1);
fmterr: /* Format error. */
if (fakeClient) freeFakeClient(fakeClient); /* avoid valgrind warning */
fclose(fp);
serverLog(LL_WARNING,"Bad file format reading the append only file: make a backup of your AOF file, then use ./redis-check-aof --fix <filename>");
exit(1);
}
......@@ -1562,39 +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 (hasActiveChildProcess()) return C_ERR;
if (aofCreatePipes() != C_OK) return C_ERR;
openChildInfoPipe();
start = ustime();
if ((childpid = fork()) == 0) {
if ((childpid = redisFork()) == 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);
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);
sendChildCOWInfo(CHILD_INFO_TYPE_AOF, "AOF rewrite");
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,
......@@ -1608,7 +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
......@@ -1623,13 +1607,14 @@ 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 (hasActiveChildProcess()) {
server.aof_rewrite_scheduled = 1;
addReplyStatus(c,"Background append only file rewriting scheduled");
} else if (rewriteAppendOnlyFileBackground() == C_OK) {
addReplyStatus(c,"Background append only file rewriting started");
} else {
addReply(c,shared.err);
addReplyError(c,"Can't execute an AOF background rewriting. "
"Please check the server logs for more information.");
}
}
......
......@@ -229,54 +229,13 @@ void disconnectAllBlockedClients(void) {
}
}
/* This function should be called by Redis every time a single command,
* a MULTI/EXEC block, or a Lua script, terminated its execution after
* being called by a client. It handles serving clients blocked in
* lists, streams, and sorted sets, via a blocking commands.
*
* All the keys with at least one client blocked that received at least
* one new element via some write operation are accumulated into
* the server.ready_keys list. This function will run the list and will
* serve clients accordingly. Note that the function will iterate again and
* again as a result of serving BRPOPLPUSH we can have new blocking clients
* to serve because of the PUSH side of BRPOPLPUSH.
*
* This function is normally "fair", that is, it will server clients
* using a FIFO behavior. However this fairness is violated in certain
* edge cases, that is, when we have clients blocked at the same time
* in a sorted set and in a list, for the same key (a very odd thing to
* do client side, indeed!). Because mismatching clients (blocking for
* a different type compared to the current key type) are moved in the
* other side of the linked list. However as long as the key starts to
* be used only for a single type, like virtually any Redis application will
* do, the function is already fair. */
void handleClientsBlockedOnKeys(void) {
while(listLength(server.ready_keys) != 0) {
list *l;
/* Point server.ready_keys to a fresh list and save the current one
* locally. This way as we run the old list we are free to call
* signalKeyAsReady() that may push new elements in server.ready_keys
* when handling clients blocked into BRPOPLPUSH. */
l = server.ready_keys;
server.ready_keys = listCreate();
while(listLength(l) != 0) {
listNode *ln = listFirst(l);
readyList *rl = ln->value;
/* First of all remove this key from db->ready_keys so that
* we can safely call signalKeyAsReady() against this key. */
dictDelete(rl->db->ready_keys,rl->key);
/* Serve clients blocked on list key. */
robj *o = lookupKeyWrite(rl->db,rl->key);
if (o != NULL && o->type == OBJ_LIST) {
dictEntry *de;
/* Helper function for handleClientsBlockedOnKeys(). This function is called
* when there may be clients blocked on a list key, and there may be new
* data to fetch (the key is ready). */
void serveClientsBlockedOnListKey(robj *o, readyList *rl) {
/* We serve clients in the same order they blocked for
* this key, from the first blocked to the last. */
de = dictFind(rl->db->blocking_keys,rl->key);
dictEntry *de = dictFind(rl->db->blocking_keys,rl->key);
if (de) {
list *clients = dictGetVal(de);
int numclients = listLength(clients);
......@@ -329,15 +288,15 @@ void handleClientsBlockedOnKeys(void) {
}
/* We don't call signalModifiedKey() as it was already called
* when an element was pushed on the list. */
}
/* Serve clients blocked on sorted set key. */
else if (o != NULL && o->type == OBJ_ZSET) {
dictEntry *de;
}
/* Helper function for handleClientsBlockedOnKeys(). This function is called
* when there may be clients blocked on a sorted set key, and there may be new
* data to fetch (the key is ready). */
void serveClientsBlockedOnSortedSetKey(robj *o, readyList *rl) {
/* We serve clients in the same order they blocked for
* this key, from the first blocked to the last. */
de = dictFind(rl->db->blocking_keys,rl->key);
dictEntry *de = dictFind(rl->db->blocking_keys,rl->key);
if (de) {
list *clients = dictGetVal(de);
int numclients = listLength(clients);
......@@ -376,10 +335,12 @@ void handleClientsBlockedOnKeys(void) {
decrRefCount(argv[1]);
}
}
}
}
/* Serve clients blocked on stream key. */
else if (o != NULL && o->type == OBJ_STREAM) {
/* Helper function for handleClientsBlockedOnKeys(). This function is called
* when there may be clients blocked on a stream key, and there may be new
* data to fetch (the key is ready). */
void serveClientsBlockedOnStreamKey(robj *o, readyList *rl) {
dictEntry *de = dictFind(rl->db->blocking_keys,rl->key);
stream *s = o->ptr;
......@@ -467,6 +428,58 @@ void handleClientsBlockedOnKeys(void) {
}
}
}
}
/* This function should be called by Redis every time a single command,
* a MULTI/EXEC block, or a Lua script, terminated its execution after
* being called by a client. It handles serving clients blocked in
* lists, streams, and sorted sets, via a blocking commands.
*
* All the keys with at least one client blocked that received at least
* one new element via some write operation are accumulated into
* the server.ready_keys list. This function will run the list and will
* serve clients accordingly. Note that the function will iterate again and
* again as a result of serving BRPOPLPUSH we can have new blocking clients
* to serve because of the PUSH side of BRPOPLPUSH.
*
* This function is normally "fair", that is, it will server clients
* using a FIFO behavior. However this fairness is violated in certain
* edge cases, that is, when we have clients blocked at the same time
* in a sorted set and in a list, for the same key (a very odd thing to
* do client side, indeed!). Because mismatching clients (blocking for
* a different type compared to the current key type) are moved in the
* other side of the linked list. However as long as the key starts to
* be used only for a single type, like virtually any Redis application will
* do, the function is already fair. */
void handleClientsBlockedOnKeys(void) {
while(listLength(server.ready_keys) != 0) {
list *l;
/* Point server.ready_keys to a fresh list and save the current one
* locally. This way as we run the old list we are free to call
* signalKeyAsReady() that may push new elements in server.ready_keys
* when handling clients blocked into BRPOPLPUSH. */
l = server.ready_keys;
server.ready_keys = listCreate();
while(listLength(l) != 0) {
listNode *ln = listFirst(l);
readyList *rl = ln->value;
/* First of all remove this key from db->ready_keys so that
* we can safely call signalKeyAsReady() against this key. */
dictDelete(rl->db->ready_keys,rl->key);
/* Serve clients blocked on list key. */
robj *o = lookupKeyWrite(rl->db,rl->key);
if (o != NULL) {
if (o->type == OBJ_LIST)
serveClientsBlockedOnListKey(o,rl);
else if (o->type == OBJ_ZSET)
serveClientsBlockedOnSortedSetKey(o,rl);
else if (o->type == OBJ_STREAM)
serveClientsBlockedOnStreamKey(o,rl);
}
/* Free this item. */
......@@ -592,7 +605,7 @@ void unblockClientWaitingData(client *c) {
* the same key again and again in the list in case of multiple pushes
* made by a script or in the context of MULTI/EXEC.
*
* The list will be finally processed by handleClientsBlockedOnLists() */
* The list will be finally processed by handleClientsBlockedOnKeys() */
void signalKeyAsReady(redisDb *db, robj *key) {
readyList *rl;
......
......@@ -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;
}
}
}
......@@ -138,6 +138,7 @@ int clusterLoadConfig(char *filename) {
/* Handle the special "vars" line. Don't pretend it is the last
* line even if it actually is when generated by Redis. */
if (strcasecmp(argv[0],"vars") == 0) {
if (!(argc % 2)) goto fmterr;
for (j = 1; j < argc; j += 2) {
if (strcasecmp(argv[j],"currentEpoch") == 0) {
server.cluster->currentEpoch =
......@@ -4251,12 +4252,9 @@ NULL
}
} else if (!strcasecmp(c->argv[1]->ptr,"nodes") && c->argc == 2) {
/* CLUSTER NODES */
robj *o;
sds ci = clusterGenNodesDescription(0);
o = createObject(OBJ_STRING,ci);
addReplyBulk(c,o);
decrRefCount(o);
sds nodes = clusterGenNodesDescription(0);
addReplyVerbatim(c,nodes,sdslen(nodes),"txt");
sdsfree(nodes);
} else if (!strcasecmp(c->argv[1]->ptr,"myid") && c->argc == 2) {
/* CLUSTER MYID */
addReplyBulkCBuffer(c,myself->name, CLUSTER_NAMELEN);
......@@ -4498,10 +4496,8 @@ NULL
"cluster_stats_messages_received:%lld\r\n", tot_msg_received);
/* Produce the reply protocol. */
addReplySds(c,sdscatprintf(sdsempty(),"$%lu\r\n",
(unsigned long)sdslen(info)));
addReplySds(c,info);
addReply(c,shared.crlf);
addReplyVerbatim(c,info,sdslen(info),"txt");
sdsfree(info);
} else if (!strcasecmp(c->argv[1]->ptr,"saveconfig") && c->argc == 2) {
int retval = clusterSaveConfig(1);
......@@ -4832,7 +4828,7 @@ int verifyDumpPayload(unsigned char *p, size_t len) {
* DUMP is actually not used by Redis Cluster but it is the obvious
* complement of RESTORE and can be useful for different applications. */
void dumpCommand(client *c) {
robj *o, *dumpobj;
robj *o;
rio payload;
/* Check if the key is here. */
......@@ -4845,9 +4841,7 @@ void dumpCommand(client *c) {
createDumpPayload(&payload,o,c->argv[1]);
/* Transfer to the client */
dumpobj = createObject(OBJ_STRING,payload.io.buffer.ptr);
addReplyBulk(c,dumpobj);
decrRefCount(dumpobj);
addReplyBulkSds(c,payload.io.buffer.ptr);
return;
}
......
This diff is collapsed.
This diff is collapsed.
......@@ -692,7 +692,8 @@ NULL
dictGetStats(buf,sizeof(buf),server.db[dbid].expires);
stats = sdscat(stats,buf);
addReplyBulkSds(c,stats);
addReplyVerbatim(c,stats,sdslen(stats),"txt");
sdsfree(stats);
} else if (!strcasecmp(c->argv[1]->ptr,"htstats-key") && c->argc == 3) {
robj *o;
dict *ht = NULL;
......@@ -719,7 +720,7 @@ NULL
} else {
char buf[4096];
dictGetStats(buf,sizeof(buf),ht);
addReplyBulkCString(c,buf);
addReplyVerbatim(c,buf,strlen(buf),"txt");
}
} else if (!strcasecmp(c->argv[1]->ptr,"change-repl-id") && c->argc == 2) {
serverLog(LL_WARNING,"Changing replication IDs after receiving DEBUG change-repl-id");
......@@ -764,7 +765,7 @@ void _serverAssertPrintClientInfo(const client *c) {
bugReportStart();
serverLog(LL_WARNING,"=== ASSERTION FAILED CLIENT CONTEXT ===");
serverLog(LL_WARNING,"client->flags = %d", c->flags);
serverLog(LL_WARNING,"client->flags = %llu", (unsigned long long)c->flags);
serverLog(LL_WARNING,"client->fd = %d", c->fd);
serverLog(LL_WARNING,"client->argc = %d", c->argc);
for (j=0; j < c->argc; j++) {
......@@ -1172,6 +1173,33 @@ void logRegisters(ucontext_t *uc) {
(unsigned long) uc->uc_mcontext.mc_cs
);
logStackContent((void**)uc->uc_mcontext.mc_rsp);
#elif defined(__aarch64__) /* Linux AArch64 */
serverLog(LL_WARNING,
"\n"
"X18:%016lx X19:%016lx\nX20:%016lx X21:%016lx\n"
"X22:%016lx X23:%016lx\nX24:%016lx X25:%016lx\n"
"X26:%016lx X27:%016lx\nX28:%016lx X29:%016lx\n"
"X30:%016lx\n"
"pc:%016lx sp:%016lx\npstate:%016lx fault_address:%016lx\n",
(unsigned long) uc->uc_mcontext.regs[18],
(unsigned long) uc->uc_mcontext.regs[19],
(unsigned long) uc->uc_mcontext.regs[20],
(unsigned long) uc->uc_mcontext.regs[21],
(unsigned long) uc->uc_mcontext.regs[22],
(unsigned long) uc->uc_mcontext.regs[23],
(unsigned long) uc->uc_mcontext.regs[24],
(unsigned long) uc->uc_mcontext.regs[25],
(unsigned long) uc->uc_mcontext.regs[26],
(unsigned long) uc->uc_mcontext.regs[27],
(unsigned long) uc->uc_mcontext.regs[28],
(unsigned long) uc->uc_mcontext.regs[29],
(unsigned long) uc->uc_mcontext.regs[30],
(unsigned long) uc->uc_mcontext.pc,
(unsigned long) uc->uc_mcontext.sp,
(unsigned long) uc->uc_mcontext.pstate,
(unsigned long) uc->uc_mcontext.fault_address
);
logStackContent((void**)uc->uc_mcontext.sp);
#else
serverLog(LL_WARNING,
" Dumping of registers not supported for this OS/arch");
......@@ -1399,6 +1427,12 @@ void sigsegvHandler(int sig, siginfo_t *info, void *secret) {
/* Log dump of processor registers */
logRegisters(uc);
/* Log Modules INFO */
serverLogRaw(LL_WARNING|LL_RAW, "\n------ MODULES INFO OUTPUT ------\n");
infostring = modulesCollectInfo(sdsempty(), NULL, 1, 0);
serverLogRaw(LL_WARNING|LL_RAW, infostring);
sdsfree(infostring);
#if defined(HAVE_PROC_MAPS)
/* Test memory */
serverLogRaw(LL_WARNING|LL_RAW, "\n------ FAST MEMORY TEST ------\n");
......
......@@ -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 (hasActiveChildProcess())
return; /* Defragging memory while there's a fork will just do damage. */
/* Once a second, check if we the fragmentation justfies starting a scan
......
......@@ -64,6 +64,7 @@ int activeExpireCycleTryExpire(redisDb *db, dictEntry *de, long long now) {
dbSyncDelete(db,keyobj);
notifyKeyspaceEvent(NOTIFY_EXPIRED,
"expired",keyobj,db->id);
trackingInvalidateKey(keyobj);
decrRefCount(keyobj);
server.stat_expiredkeys++;
return 1;
......
......@@ -466,7 +466,7 @@ void georadiusGeneric(client *c, int flags) {
/* Look up the requested zset */
robj *zobj = NULL;
if ((zobj = lookupKeyReadOrReply(c, key, shared.null[c->resp])) == NULL ||
if ((zobj = lookupKeyReadOrReply(c, key, shared.emptyarray)) == NULL ||
checkType(c, zobj, OBJ_ZSET)) {
return;
}
......@@ -566,7 +566,7 @@ void georadiusGeneric(client *c, int flags) {
/* If no matching results, the user gets an empty reply. */
if (ga->used == 0 && storekey == NULL) {
addReplyNull(c);
addReply(c,shared.emptyarray);
geoArrayFree(ga);
return;
}
......
......@@ -700,7 +700,7 @@ int hllSparseSet(robj *o, long index, uint8_t count) {
p += oplen;
first += span;
}
if (span == 0) return -1; /* Invalid format. */
if (span == 0 || p >= end) return -1; /* Invalid format. */
next = HLL_SPARSE_IS_XZERO(p) ? p+2 : p+1;
if (next >= end) next = NULL;
......@@ -1242,7 +1242,7 @@ void pfcountCommand(client *c) {
if (o == NULL) continue; /* Assume empty HLL for non existing var.*/
if (isHLLObjectOrReply(c,o) != C_OK) return;
/* Merge with this HLL with our 'max' HHL by setting max[i]
/* Merge with this HLL with our 'max' HLL by setting max[i]
* to MAX(max[i],hll[i]). */
if (hllMerge(registers,o) == C_ERR) {
addReplySds(c,sdsnew(invalid_hll_err));
......@@ -1329,7 +1329,7 @@ void pfmergeCommand(client *c) {
hdr = o->ptr;
if (hdr->encoding == HLL_DENSE) use_dense = 1;
/* Merge with this HLL with our 'max' HHL by setting max[i]
/* Merge with this HLL with our 'max' HLL by setting max[i]
* to MAX(max[i],hll[i]). */
if (hllMerge(max,o) == C_ERR) {
addReplySds(c,sdsnew(invalid_hll_err));
......
......@@ -599,7 +599,7 @@ NULL
event = dictGetKey(de);
graph = latencyCommandGenSparkeline(event,ts);
addReplyBulkCString(c,graph);
addReplyVerbatim(c,graph,sdslen(graph),"txt");
sdsfree(graph);
} else if (!strcasecmp(c->argv[1]->ptr,"latest") && c->argc == 2) {
/* LATENCY LATEST */
......@@ -608,7 +608,7 @@ NULL
/* LATENCY DOCTOR */
sds report = createLatencyReport();
addReplyBulkCBuffer(c,report,sdslen(report));
addReplyVerbatim(c,report,sdslen(report),"txt");
sdsfree(report);
} else if (!strcasecmp(c->argv[1]->ptr,"reset") && c->argc >= 2) {
/* LATENCY RESET */
......
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
This diff is collapsed.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment