github valkey-io/valkey 9.0.8

43 minutes ago

Valkey 9.0.8 - Released Wed 18 November 2026

Upgrade urgency SECURITY: This release includes security fixes we recommend you apply as soon as possible.

Security Fixes

  • CVE-2026-130822: Fix a use-after-free crash in client eviction when maxmemory-clients is enabled by @nadav-levanoni (#5435)

Bug Fixes

  • Fix a server crash when a key with a name of 2 GiB or more is sampled for eviction, and hangs, crashes, or duplicate set members when comparing strings of that size by @madolson (#5133)
  • Correct command metadata to describe the ZINCRBY increment and ZINTER/ZINTERSTORE/ZUNION/ZUNIONSTORE weights as doubles instead of integers by @sarthakaggarwal97 (#5210)
  • Propagate PFADD elements applied before a corrupted-HLL INVALIDOBJ error so replicas and the AOF stay consistent with the primary by @madolson (#5128)
  • Correct expiry command metadata to show that XX can be combined with GT or LT, rather than all four conditions being mutually exclusive by @sarthakaggarwal97 (#5213)
  • Fix a server panic when a Lua script returns or raises an error table whose metatable has an __index metamethod that raises an error by @madolson (#5118)
  • Fix a server crash when a module never sets a postponed reply length; the leftover placeholder is now filled with an empty array by @madolson (#5159)
  • Mark the RESTORE and RESTORE-ASKING IDLETIME and FREQ options as mutually exclusive in command metadata by @sarthakaggarwal97 (#5212)
  • Fix a crash when processing a heartbeat from a cluster node whose role is locally unknown, such as after loading an incomplete nodes.conf by @enjoy-binbin (#4091)
  • Fix a memory leak when server.register_function named arguments are repeated in a different letter case by @madolson (#5120)
  • Fix quadratic slowdown in HRANDFIELD and SRANDMEMBER with counts on collections whose elements differ only in letter case by @madolson (#5167)
  • Fix a crash in a cluster replica whose primary was removed from the nodes table when another primary claims an unassigned slot by @madolson (#5184)
  • Fix SCRIPT KILL and FUNCTION KILL not stopping Lua scripts that loop inside nested coroutines, and script timeouts not being detected when work is spread across coroutines by @madolson (#5108)
  • Fix an out-of-bounds write in PFADD and PFMERGE on a corrupted sparse HyperLogLog ending in a truncated XZERO opcode by rejecting it as invalid by @madolson (#5124)
  • Fix a server abort in PFADD and PFMERGE on short sparse HyperLogLogs when hll-sparse-max-bytes is set to 31 or less by @madolson (#5138)
  • Fix a server crash in SORT and SORT_RO with BY or GET patterns when an element is 2 GiB or larger by @madolson (#5146)
  • Fix ValkeyModule_Calloc and ValkeyModule_TryCalloc returning an undersized buffer when nmemb * size overflows by @madolson (#5155)
  • Fix ValkeyModule_CallReplyAttributeElement always returning an error on RESP3 attribute replies by @madolson (#5157)
  • Fix EXEC succeeding after a key that was already expired at WATCH time is overwritten, which can happen on replicas by @madolson (#5169)
  • Correct command metadata so the numfields count argument of hash field commands is no longer associated with a key specification by @sarthakaggarwal97 (#5211)
  • Fix PFMERGE with no source keys bypassing ACL key permissions and failing in COMMAND GETKEYS, and a crash in module commands whose keynum key spec starts past the last argument by @madolson (#5087)
  • Fix HSETEX with a repeated EX, PX or EXAT option being rejected by replicas and lost on AOF reload by @madolson (#5088)
  • Fix a memory leak when a module command is refused by ACL while the ACL log is disabled with acllog-max-len 0 by @madolson (#5162)
  • Fix a crash on the first FUNCTION LOAD after startup when the shebang engine name contains a NUL byte by @madolson (#5115)
  • Fix an active defrag crash when a large set, sorted set or list is replaced by a hashtable-encoded hash while it is being defragmented by @madolson (#5164)
  • Fix FCALL errors or a crash when FUNCTION FLUSH SYNC runs before the lazy free triggered by FUNCTION RESTORE of more than 64 functions has completed by @madolson (#5105)
  • Fix long stalls in SRANDMEMBER, ZRANDMEMBER, HRANDFIELD and SPOP on a small key whose hashtable is still shrinking from a much larger size by @madolson (#5170)
  • Fix notify-keyspace-events dropping the n (new key) flag from CONFIG GET and CONFIG REWRITE when A is also set by @pirate329 (#5199)
  • Fix a cluster node failing to restart with an assertion after a gossip entry carrying the MYSELF flag adds a second myself node to nodes.conf by @madolson (#5134)
  • Fix ACL LOAD not disconnecting default user clients subscribed to channels the reloaded rules no longer allow by @madolson (#5085)
  • Fix active defrag crash when a database's keyspace is replaced (async FLUSHDB/FLUSHALL, SWAPDB, replica full sync) while a large key is being defragmented by @taham0 (#4968)
  • Allow modules to block RESP3 clients that are subscribed to Pub/Sub channels instead of failing ValkeyModule_BlockClient with ENOTSUP by @quanyeyang (#4863)
  • Bound the time spent per cycle evicting keys from the client tracking table, avoiding long stalls after lowering tracking-table-max-keys by @tanmayBeginsJourney (#4775)
  • Fix a memory leak when GEOSEARCH or GEOSEARCHSTORE is given repeated BYPOLYGON options by @charsyam (#4995)
  • MEMORY USAGE now counts hash fields that have expired but not yet been reaped, fixing under-reported sizes by @sushilpaneru1 (#5092)
  • Fix a memory leak in ValkeyModule_GetCommandKeysWithFlags when introspecting commands with large key lists by @christopherlim98 (#5046)
  • In cluster mode, one cross-slot command in a script no longer makes every later script and function call fail until restart by @madolson (#5191)
  • Correct SENTINEL SIMULATE-FAILURE metadata so COMMAND DOCS shows the required mode argument and its tokens by @sarthakaggarwal97 (#5208)
  • Fix a TLS busy loop that consumed 100% CPU when a connection's read or write handler was removed while OpenSSL was waiting for the opposite event by @yairgott (#3510)
  • Fix a crash caused by a NULL pointer dereference when all timer events in an event loop are pending deletion by @royenheart (#4351)
  • Fix a node staying stuck in PFAIL when its PING is lost while the peer's own traffic keeps the cluster link alive by @enjoy-binbin (#4171)
  • CLUSTER SYNCSLOTS FINISH now rejects clients outside the slot migration stream, preventing misuse of the import state machine by @enjoy-binbin (#4330)
  • Reject a repeated CLUSTER SYNCSLOTS ESTABLISH on the same connection, which could lead to a use-after-free crash on disconnect by @Pyolar (#4670)
  • Fix SORT key extraction so a STORE destination named like an option can no longer bypass ACL checks and write to an unauthorized key by @madolson (#4665)
  • Fix GEOSEARCH returning empty results for exact zero-distance matches on builds where floating-point contraction breaks the latitude distance calculation by @Baraa-Hasheesh (#4777)
  • Fix a crash when a replication role change command such as REPLICAOF or CLUSTER REPLICATE arrives over a replication link by @madolson (#5187)
  • Fix a crash on the importing primary when an atomic slot migration fails while its writes are paused by @madolson (#5182)
  • Fix XADD with elements over the size limit leaving an empty stream on the primary that is never propagated to replicas or the AOF by @madolson (#5178)
  • Redact credentials of rejected HELLO, AUTH and MIGRATE commands so passwords no longer leak to MONITOR and the commandlog by @madolson (#5180)
  • Prevent a replica crash when the primary wrongly replies +CONTINUE to a full resync request; the replica now drops the link and retries by @madolson (#5126)
  • Fix an integer overflow on 64-bit builds that crashes the server when merging list nodes whose combined size exceeds 4GiB, reachable through a RESTORE or full sync payload followed by LSET by @madolson (#5163)
  • Fix a stack buffer overflow when replying with hundreds of empty strings (e.g. MGET) while reply copy avoidance is active by @madolson (#5151)
  • Ignore invalid ping extension values from cluster peers that would corrupt nodes.conf and prevent the node from restarting by @madolson (#5137)
  • Reclaim client-side caching tracking table entries left by disconnected clients, fixing a memory leak under read-heavy workloads, and include the tracking table in active defrag by @rayjinghaolei (#4190)
  • Bound the bytes the cluster bus reads per event so a peer with a constant read backlog cannot starve clients, cron and heartbeats by @madolson (#5161)
  • Mark the absolute timestamp arguments of HEXPIREAT and HPEXPIREAT as unix-time in COMMAND DOCS, matching EXPIREAT and PEXPIREAT by @rainsupreme (#5245)
  • Correct ZADD and LCS metadata so COMMAND DOCS and valkey-cli hints no longer present NX with GT/LT or LEN with IDX as valid combinations by @rainsupreme (#5244)
  • Fail CONFIG REWRITE when the old config file reads short instead of writing uninitialized memory into the new config file by @madolson (#5158)
  • Restore acceptance of hexadecimal floats (such as 0x10) in sorted set score ranges and SORT when built with USE_FAST_FLOAT=yes by @quanyeyang (#4875)
  • Fix a Sentinel crash when a monitored replica's INFO reply contains no master_host by @smkher (#4929)
  • Fix a use-after-free in streams when XADD creates a new node, which could leave a dangling pointer hit by later reads such as XRANGE by @Baraa-Hasheesh (#4711)
  • Reject inbound cluster bus packets larger than 16MB, except Pub/Sub and module messages, so a peer cannot make a node buffer up to 4GB from a single header by @madolson (#5065)
  • Fix XINFO GROUPS and XINFO STREAM ... FULL reporting too-high consumer group lag after XTRIM or a trimming XADD removes unread entries by @Sasidharan3094 (#4983)
  • Fix shutdown-triggered failover promoting a replica that has cluster-replica-no-failover yes by @simonasr (#5033)
  • Fix AOF load failure after a crash leaves a torn annotation at the end of the file; it is now treated as a short read and can be truncated by aof-load-truncated or valkey-check-aof --fix by @sumitk163 (#5349)
  • Correct the SENTINEL GET-PRIMARY-ADDR-BY-NAME reply schema to allow the null reply returned for an unknown primary name by @rainsupreme (#5246)
  • Fix a crash after REPLICAOF NO ONE with TLS and a connection leak after REPLICAOF to a new primary when the old primary client is freed asynchronously due to pending I/O by @yaronsananes (#3719)

Full Changelog: 9.0.7...9.0.8

Don't miss a new valkey release

NewReleases is sending notifications on new releases.