With dynamically growing argc (#9528), it is necessary to initialize
argv_len. Normally createClient() handles that, but in the case of a
module shared_client, this needs to be done explicitly.
This also addresses an issue with rewriteClientCommandArgument() which
doesn't properly handle the case where the new element extends beyond
argc but not beyond argv_len.
LCS can allocate immense amount of memory (sizes of two inputs multiplied by each other).
In the past this caused some possible security issues due to overflows, which we solved
and also added use of `trymalloc` to return "Insufficient memory" instead of OOM panic zmalloc.
But in case overcommit is enabled, it could be that we won't get the OOM panic, and zmalloc
will succeed, and then we can get OOM killed by the kernel.
The solution here is to prevent LCS from allocating transient memory that's bigger than
`proto-max-bulk-len` config.
This config is not directly related to transient memory, but using a hard coded value ad well as
introducing a specific config seems wrong.
This comes to solve an error in the corrupt-dump-fuzzer test that started in the daily CI see #9799
Recently we started using list-compress-depth in tests (was completely untested till now).
Turns this triggered test failures with the external mode, since the tests left the setting enabled
and then it was used in other tests (specifically the fuzzer named "Stress tester for #3343-alike bugs").
This PR fixes the issue of the `recompress` flag being left set by mistake, which caused the code to
later to compress the head or tail nodes (which should never be compressed)
The solution is to reset the recompress flag when it should have been (when it was decided not to compress).
Additionally we're adding some assertions and improve the tests so in order to catch other similar bugs.
Currently PING returns different status when server is not serving data,
for example when `LOADING` or `BUSY`.
But same was not true for `MASTERDOWN`
This commit makes PING reply with `MASTERDOWN` when
replica-serve-stale-data=no and link is MASTER is down.
Drop the STRALGO command, now LCS is a command of its own and it only works on keys (not input strings).
The motivation is that STRALGO's syntax was really messed-up...
- assumes all (future) string algorithms will take similar arguments
- mixes command that takes keys and one that doesn't in the same command.
- make it nearly impossible to expose the right key spec in COMMAND INFO (issues cluster clients)
- hard for cluster clients to determine the key names (firstkey, lastkey, etc)
- hard for ACL / flags (is it a read command?)
This is a breaking change.
Moves ZPOP ... 0 fast exit path after type check to reply with
WRONGTYPE. In the past it will return an empty array.
Also now count is not allowed to be negative.
see #9680
before:
```
127.0.0.1:6379> set zset str
OK
127.0.0.1:6379> zpopmin zset 0
(empty array)
127.0.0.1:6379> zpopmin zset -1
(empty array)
```
after:
```
127.0.0.1:6379> set zset str
OK
127.0.0.1:6379> zpopmin zset 0
(error) WRONGTYPE Operation against a key holding the wrong kind of value
127.0.0.1:6379> zpopmin zset -1
(error) ERR value is out of range, must be positive
```
in `scan 0 match ""` case, pat is empty sds(patlen is 0), I don't think should access the
first character directly in this case(even though the first character is ' \0 '), for the
code readability, I switch the two positions of judgment logic.
Redis supports inserting data over 4GB into string (and recently for lists too, see #9357),
But LZF compression used in RDB files (see `rdbcompression` config), and in quicklist
(see `list-compress-depth` config) does not support compress/decompress data over
UINT32_MAX, which will result in corrupting the rdb after compression.
Internal changes:
1. Modify the `unsigned int` parameter of `lzf_compress/lzf_decompress` to `size_t`.
2. Modify the variable types in `lzf_compress` involving offsets and lengths to `size_t`.
3. Set LZF_USE_OFFSETS to 0.
When LZF_USE_OFFSETS is 1, lzf store offset into `LZF_HSLOT`(32bit).
Even in 64-bit, `LZF_USE_OFFSETS` defaults to 1, because lzf assumes that it only
compresses and decompresses data smaller than UINT32_MAX.
But now we need to make lzf support 64-bit, turning on `LZF_USE_OFFSETS` will make
it impossible to store 64-bit offsets or pointers.
BTW, disable LZF_USE_OFFSETS also brings a few performance improvements.
Tests:
1. Add test for compress/decompress string large than UINT32_MAX.
2. Add unittest for compress/decompress quicklistNode.
The client flags is a 64 bit integer, but the temporary cached value on the stack of call() is 32 bit.
luckily this doesn't lead to any bugs since the only flags used against this variables are below 32 bit.
- Added sanitizer support. `address`, `undefined` and `thread` sanitizers are available.
- To build Redis with desired sanitizer : `make SANITIZER=undefined`
- There were some sanitizer findings, cleaned up codebase
- Added tests with address and undefined behavior sanitizers to daily CI.
- Added tests with address sanitizer to the per-PR CI (smoke out mem leaks sooner).
Basically, there are three types of issues :
**1- Unaligned load/store** : Most probably, this issue may cause a crash on a platform that
does not support unaligned access. Redis does unaligned access only on supported platforms.
**2- Signed integer overflow.** Although, signed overflow issue can be problematic time to time
and change how compiler generates code, current findings mostly about signed shift or simple
addition overflow. For most platforms Redis can be compiled for, this wouldn't cause any issue
as far as I can tell (checked generated code on godbolt.org).
**3 -Minor leak** (redis-cli), **use-after-free**(just before calling exit());
UB means nothing guaranteed and risky to reason about program behavior but I don't think any
of the fixes here worth backporting. As sanitizers are now part of the CI, preventing new issues
will be the real benefit.
During diskless replication, the check for broken EOF mark is misplaced
and should be earlier. Now we do not swap db, we do proper cleanup and
correctly raise module events on this kind of failure.
This issue existed prior to #9323, but before, the side effect was not restoring
backup and not raising the correct module events on this failure.
* Clean up EINTR handling so EINTR will not change connection state to begin with.
* On TLS, catch EINTR and return it as-is before going through OpenSSL error handling (which seems to not distinguish it from EAGAIN).
This refactors all `CONFIG SET`s and conf file loading arguments go through
the generic config handling interface.
Refactoring changes:
- All config params go through the `standardConfig` interface (some stuff which
is only related to the config file and not the `CONFIG` command still has special
handling for rewrite/config file parsing, `loadmodule`, for example.) .
- Added `MULTI_ARG_CONFIG` flag for configs to signify they receive a variable
number of arguments instead of a single argument. This is used to break up space
separated arguments to `CONFIG SET` so the generic setter interface can pass
multiple arguments to the setter function. When parsing the config file we also break
up anything after the config name into multiple arguments to the setter function.
Interface changes:
- A side effect of the above interface is that the `bind` argument in the config file can
be empty (no argument at all) this is treated the same as passing an single empty
string argument (same as `save` already used to work).
- Support rewrite and setting `watchdog-period` from config file (was only supported
by the CONFIG command till now).
- Another side effect is that the `save T X` config argument now supports multiple
Time-Changes pairs in a single line like its `CONFIG SET` counterpart. So in the
config file you can either do:
```
save 3600 1
save 600 10
```
or do
```
save 3600 1 600 10
```
Co-authored-by: Bjorn Svensson <bjorn.a.svensson@est.tech>
For diskless replication in swapdb mode, considering we already spend replica memory
having a backup of current db to restore in case of failure, we can have the following benefits
by instead swapping database only in case we succeeded in transferring db from master:
- Avoid `LOADING` response during failed and successful synchronization for cases where the
replica is already up and running with data.
- Faster total time of diskless replication, because now we're moving from Transfer + Flush + Load
time to Transfer + Load only. Flushing the tempDb is done asynchronously after swapping.
- This could be implemented also for disk replication with similar benefits if consumers are willing
to spend the extra memory usage.
General notes:
- The concept of `backupDb` becomes `tempDb` for clarity.
- Async loading mode will only kick in if the replica is syncing from a master that has the same
repl-id the one it had before. i.e. the data it's getting belongs to a different time of the same timeline.
- New property in INFO: `async_loading` to differentiate from the blocking loading
- Slot to Key mapping is now a field of `redisDb` as it's more natural to access it from both server.db
and the tempDb that is passed around.
- Because this is affecting replicas only, we assume that if they are not readonly and write commands
during replication, they are lost after SYNC same way as before, but we're still denying CONFIG SET
here anyways to avoid complications.
Considerations for review:
- We have many cases where server.loading flag is used and even though I tried my best, there may
be cases where async_loading should be checked as well and cases where it shouldn't (would require
very good understanding of whole code)
- Several places that had different behavior depending on the loading flag where actually meant to just
handle commands coming from the AOF client differently than ones coming from real clients, changed
to check CLIENT_ID_AOF instead.
**Additional for Release Notes**
- Bugfix - server.dirty was not incremented for any kind of diskless replication, as effect it wouldn't
contribute on triggering next database SAVE
- New flag for RM_GetContextFlags module API: REDISMODULE_CTX_FLAGS_ASYNC_LOADING
- Deprecated RedisModuleEvent_ReplBackup. Starting from Redis 7.0, we don't fire this event.
Instead, we have the new RedisModuleEvent_ReplAsyncLoad holding 3 sub-events: STARTED,
ABORTED and COMPLETED.
- New module flag REDISMODULE_OPTIONS_HANDLE_REPL_ASYNC_LOAD for RedisModule_SetModuleOptions
to allow modules to declare they support the diskless replication with async loading (when absent, we fall
back to disk-based loading).
Co-authored-by: Eduardo Semprebon <edus@saxobank.com>
Co-authored-by: Oran Agra <oran@redislabs.com>
Introduced in #8179, this fixes the command's replies in the 0 count edge case.
[BREAKING] changes the reply type when count is 0 to an empty array (instead of nil)
Moves LPOP ... 0 fast exit path after type check to reply with WRONGTYPE
When repl-diskless-load is enabled, the connection is set to the blocking state.
The connection may be interrupted by a signal during a system call.
This would have resulted in a disconnection and possibly a reconnection loop.
Co-authored-by: Oran Agra <oran@redislabs.com>
Redis lists are stored in quicklist, which is currently a linked list of ziplists.
Ziplists are limited to storing elements no larger than 4GB, so when bigger
items are added they're getting truncated.
This PR changes quicklists so that they're capable of storing large items
in quicklist nodes that are plain string buffers rather than ziplist.
As part of the PR there were few other changes in redis:
1. new DEBUG sub-commands:
- QUICKLIST-PACKED-THRESHOLD - set the threshold of for the node type to
be plan or ziplist. default (1GB)
- QUICKLIST <key> - Shows low level info about the quicklist encoding of <key>
2. rdb format change:
- A new type was added - RDB_TYPE_LIST_QUICKLIST_2 .
- container type (packed / plain) was added to the beginning of the rdb object
(before the actual node list).
3. testing:
- Tests that requires over 100MB will be by default skipped. a new flag was
added to 'runtest' to run the large memory tests (not used by default)
Co-authored-by: sundb <sundbcn@gmail.com>
Co-authored-by: Oran Agra <oran@redislabs.com>
Add new no-mandatory-keys flag to support COMMAND GETKEYS of commands
which have no mandatory keys.
In the past we would have got this error:
```
127.0.0.1:6379> command getkeys eval "return 1" 0
(error) ERR Invalid arguments specified for command
```
When using SETNX and SETXX we could end up doing key lookup twice.
This presents a small inefficiency price.
Also once we have statistics of write hit and miss they'll be wrong (recording the same key hit twice)
Since the loop in incrementalTrimReplicationBacklog checks the size of histlen,
we cannot afford to update it only when the loop exits, this may cause deleting
much more replication blocks, and replication backlog may be less than setting size.
introduce in #9166
Co-authored-by: sundb <sundbcn@gmail.com>
After PR #9166 , replication backlog is not a real block of memory, just contains a
reference points to replication buffer's block and the blocks index (to accelerate
search offset when partial sync), so we need update both replication buffer's block's
offset and replication backlog blocks index's offset when master restart from RDB,
since the `server.master_repl_offset` is changed.
The implications of this bug was just a slow search, but not a replication failure.
The module test in reply.tcl was introduced by #8521 but didn't run until recently (see #9639)
and then it started failing with valgrind.
This is because valgrind uses 64 bit long double (unlike most other platforms that have at least 80 bits)
But besides valgrind, the tests where also incompatible with ARM32, which also uses 64 bit long doubles.
We now use appropriate value to avoid issues with either valgrind or ARM32
In all the double tests, i use 3.141, which is safe since since addReplyDouble uses
`%.17Lg` which is able to represent this value without adding any digits due to precision loss.
In the long double, since we use `%.17Lf` in ld2string, it preserves 17 significant
digits, rather than 17 digit after the decimal point (like in `%.17Lg`).
So to make these similar, i use value lower than 1 (no digits left of
the period)
Lastly, we have the same issue with TCL (no long doubles) so we read
raw protocol in that test.
Note that the only error before this fix (in both valgrind and ARM32 is this:
```
*** [err]: RM_ReplyWithLongDouble: a float reply in tests/unit/moduleapi/reply.tcl
Expected '3.141' to be equal to '3.14100000000000001' (context: type eval line 2 cmd {assert_equal 3.141 [r rw.longdouble 3.141]} proc ::test)
```
so the changes to debug.c and scripting.tcl aren't really needed, but i consider them a cleanup
(i.e. scripting.c validated a different constant than the one that's sent to it from debug.c).
Another unrelated change is to add the RESP version to the repeated tests in reply.tcl
The previous code did not check whether COUNT is set.
So we can use `lmpop 2 key1 key2 left count 1 count 2`.
This situation can occur in LMPOP/BLMPOP/ZMPOP/BZMPOP commands.
LMPOP/BLMPOP introduced in #9373, ZMPOP/BZMPOP introduced in #9484.
Add timestamp annotation in AOF, one part of #9325.
Enabled with the new `aof-timestamp-enabled` config option.
Timestamp annotation format is "#TS:${timestamp}\r\n"."
TS" is short of timestamp and this method could save extra bytes in AOF.
We can use timestamp annotation for some special functions.
- know the executing time of commands
- restore data to a specific point-in-time (by using redis-check-rdb to truncate the file)
Improve code doc for allowed_firstargs (used to be allowed_commands before #9504.
I don't think the text in the code needs to refer to the history (it's not there just for backwards compatibility).
instead it should just describe what it does.
Let modules use additional type of RESP3 response (unused by redis so far)
Also fix tests that where introduced in #8521 but didn't actually run.
Co-authored-by: Oran Agra <oran@redislabs.com>
## Background
For redis master, one replica uses one copy of replication buffer, that is a big waste of memory,
more replicas more waste, and allocate/free memory for every reply list also cost much.
If we set client-output-buffer-limit small and write traffic is heavy, master may disconnect with
replicas and can't finish synchronization with replica. If we set client-output-buffer-limit big,
master may be OOM when there are many replicas that separately keep much memory.
Because replication buffers of different replica client are the same, one simple idea is that
all replicas only use one replication buffer, that will effectively save memory.
Since replication backlog content is the same as replicas' output buffer, now we
can discard replication backlog memory and use global shared replication buffer
to implement replication backlog mechanism.
## Implementation
I create one global "replication buffer" which contains content of replication stream.
The structure of "replication buffer" is similar to the reply list that exists in every client.
But the node of list is `replBufBlock`, which has `id, repl_offset, refcount` fields.
```c
/* Replication buffer blocks is the list of replBufBlock.
*
* +--------------+ +--------------+ +--------------+
* | refcount = 1 | ... | refcount = 0 | ... | refcount = 2 |
* +--------------+ +--------------+ +--------------+
* | / \
* | / \
* | / \
* Repl Backlog Replia_A Replia_B
*
* Each replica or replication backlog increments only the refcount of the
* 'ref_repl_buf_node' which it points to. So when replica walks to the next
* node, it should first increase the next node's refcount, and when we trim
* the replication buffer nodes, we remove node always from the head node which
* refcount is 0. If the refcount of the head node is not 0, we must stop
* trimming and never iterate the next node. */
/* Similar with 'clientReplyBlock', it is used for shared buffers between
* all replica clients and replication backlog. */
typedef struct replBufBlock {
int refcount; /* Number of replicas or repl backlog using. */
long long id; /* The unique incremental number. */
long long repl_offset; /* Start replication offset of the block. */
size_t size, used;
char buf[];
} replBufBlock;
```
So now when we feed replication stream into replication backlog and all replicas, we only need
to feed stream into replication buffer `feedReplicationBuffer`. In this function, we set some fields of
replication backlog and replicas to references of the global replication buffer blocks. And we also
need to check replicas' output buffer limit to free if exceeding `client-output-buffer-limit`, and trim
replication backlog if exceeding `repl-backlog-size`.
When sending reply to replicas, we also need to iterate replication buffer blocks and send its
content, when totally sending one block for replica, we decrease current node count and
increase the next current node count, and then free the block which reference is 0 from the
head of replication buffer blocks.
Since now we use linked list to manage replication backlog, it may cost much time for iterating
all linked list nodes to find corresponding replication buffer node. So we create a rax tree to
store some nodes for index, but to avoid rax tree occupying too much memory, i record
one per 64 nodes for index.
Currently, to make partial resynchronization as possible as much, we always let replication
backlog as the last reference of replication buffer blocks, backlog size may exceeds our setting
if slow replicas that reference vast replication buffer blocks, and this method doesn't increase
memory usage since they share replication buffer. To avoid freezing server for freeing unreferenced
replication buffer blocks when we need to trim backlog for exceeding backlog size setting,
we trim backlog incrementally (free 64 blocks per call now), and make it faster in
`beforeSleep` (free 640 blocks).
### Other changes
- `mem_total_replication_buffers`: we add this field in INFO command, it means the total
memory of replication buffers used.
- `mem_clients_slaves`: now even replica is slow to replicate, and its output buffer memory
is not 0, but it still may be 0, since replication backlog and replicas share one global replication
buffer, only if replication buffer memory is more than the repl backlog setting size, we consider
the excess as replicas' memory. Otherwise, we think replication buffer memory is the consumption
of repl backlog.
- Key eviction
Since all replicas and replication backlog share global replication buffer, we think only the
part of exceeding backlog size the extra separate consumption of replicas.
Because we trim backlog incrementally in the background, backlog size may exceeds our
setting if slow replicas that reference vast replication buffer blocks disconnect.
To avoid massive eviction loop, we don't count the delayed freed replication backlog into
used memory even if there are no replicas, i.e. we also regard this memory as replicas's memory.
- `client-output-buffer-limit` check for replica clients
It doesn't make sense to set the replica clients output buffer limit lower than the repl-backlog-size
config (partial sync will succeed and then replica will get disconnected). Such a configuration is
ignored (the size of repl-backlog-size will be used). This doesn't have memory consumption
implications since the replica client will share the backlog buffers memory.
- Drop replication backlog after loading data if needed
We always create replication backlog if server is a master, we need it because we put DELs in
it when loading expired keys in RDB, but if RDB doesn't have replication info or there is no rdb,
it is not possible to support partial resynchronization, to avoid extra memory of replication backlog,
we drop it.
- Multi IO threads
Since all replicas and replication backlog use global replication buffer, if I/O threads are enabled,
to guarantee data accessing thread safe, we must let main thread handle sending the output buffer
to all replicas. But before, other IO threads could handle sending output buffer of all replicas.
## Other optimizations
This solution resolve some other problem:
- When replicas disconnect with master since of out of output buffer limit, releasing the output
buffer of replicas may freeze server if we set big `client-output-buffer-limit` for replicas, but now,
it doesn't cause freezing.
- This implementation may mitigate reply list copy cost time(also freezes server) when one replication
has huge reply buffer and another replica can copy buffer for full synchronization. now, we just copy
reference info, it is very light.
- If we set replication backlog size big, it also may cost much time to copy replication backlog into
replica's output buffer. But this commit eliminates this problem.
- Resizing replication backlog size doesn't empty current replication backlog content.
I moved a bunch of stats in redisFork to be executed only on successful
fork, since they seem wrong to be done when it failed.
I guess when fork fails it does that immediately, no latency spike.
Before this commit, module blocked clients did not carry through the original RESP version, resulting with RESP3 clients receiving unexpected RESP2 replies.
