# \[RELEASE\] ScyllaDB 2025.4.2 **URL:** \ **Category:** Release Notes\ **Tags:** enterprise, enterprise-release, scylladb-2025-4\ **Created:** [January 26, 2026, 10:51am UTC](https://forum.scylladb.com/t/release-scylladb-2025-4-2/5251 "2026-01-26T10:51:47Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![Anna](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.scylladb.com/anna/32/48_2.png) [@Anna](https://forum.scylladb.com/u/Anna)\ **Post date:** [January 26, 2026, 10:51am UTC](https://forum.scylladb.com/t/release-scylladb-2025-4-2/5251/1 "2026-01-26T10:51:47Z")
The ScyllaDB team is pleased to announce the release of ScyllaDB 2025.4.2, a production-ready patch release for [ScyllaDB 2025.4 Feature Release](https://forum.scylladb.com/t/release-scylladb-2025-4-0/5222). ## Related Links - [Get ScyllaDB 2025.4](https://docs.scylladb.com/manual/branch-2025.4/getting-started/install-scylla/index.html) - [Upgrade ScyllaDB 2025.4 (Patch Upgrade)](https://docs.scylladb.com/manual/branch-2025.4/upgrade/upgrade-guides/upgrade-guide-from-2025.4.x-to-2025.4.y.html#upgrade-scylla-name-src-version-to-new-version-patch-upgrades) - [Upgrade from ScyllaDB 2025.x to ScyllaDB 2025.4](https://docs.scylladb.com/manual/branch-2025.4/upgrade/upgrade-guides/upgrade-guide-from-2025.x-to-2025.4/index.html) - [Contact our Support Team](https://www.scylladb.com/product/support/#enterprise-support) ## Bug Fixes The following issues are fixed in this release. ### **Alternator** - Alternator LWT Batch Write Fix: The system now continuously monitors the tablet’s location during the entire batch write process. If the tablet moves to a new processing unit, the write is automatically and immediately rerouted to the correct new location, ensuring the write operation completes successfully and maintaining full data consistency for LWTs during active tablet migrations. Additionally, memory stability and efficiency were enhanced by preventing accidental cross-shard memory deallocation and reducing memory usage by eliminating redundant copies of internal key data created during certain operations. [scylladb#27529](https://github.com/scylladb/scylladb/pull/27529) ### **Authentication & Authorization** - Added exponential retry for server errors in the Google Cloud Platform (GCP) encryption key provider, increasing the resilience and reliability of key retrieval during temporary service disruptions. [scylladb#27242](https://github.com/scylladb/scylladb/issues/27242), [scylladb#27443](https://github.com/scylladb/scylladb/pull/27443) ### **Vector Search** - Resolved a bug that was causing failures in Change Data Capture (CDC) vector search indexing due to incorrect permission handling. [scylladb#28068](https://github.com/scylladb/scylladb/pull/28068) - Granted system table permissions to the **VECTOR\_SEARCH\_INDEXING** role. [scylladb#28042](https://github.com/scylladb/scylladb/pull/28042) ### **Commitlog** - Fixed a race condition between large allocation and segment termination. [scylladb#27992](https://github.com/scylladb/scylladb/issues/27992), [scylladb#28099](https://github.com/scylladb/scylladb/pull/28099) - Fixed a bug in log file reading to check for end-of-file position on all data reads. [scylladb#27346](https://github.com/scylladb/scylladb/pull/27346) ### **Batchtlog** - Improved batchlog management to only update the last replay marker if all batches were successfully replayed, and corrected an issue where failed batches could resurrect tombstone-garbage collected data. [scylladb#24415](https://github.com/scylladb/scylladb/issues/24415), [scylladb#27094](https://github.com/scylladb/scylladb/pull/27094) ### **Repair** - Corrected the range format for repair history to ensure it adheres to the expected (start, end) tuple structure, which improves data consistency and accuracy for repair operations. [scylladb#27220](https://github.com/scylladb/scylladb/issues/27220), [scylladb#27461](https://github.com/scylladb/scylladb/pull/27461) ### **Hints** - Prevented draining hints before hint replay is allowed to ensure correctness, safeguarding against potential data inconsistencies or loss of hinted writes. [scylladb#27972](https://github.com/scylladb/scylladb/pull/27972) ### **Raft & Topology** - Preserved the IP-to-ID mapping of a replacing node on restart, which is crucial for maintaining the correct cluster topology and ensuring a smooth, consistent replacement process without data inconsistencies or communication issues. [scylladb#28057](https://github.com/scylladb/scylladb/issues/28057), [scylladb#28100](https://github.com/scylladb/scylladb/pull/28100) - Added a barrier to the **cleanup\_target** in the topology coordinator to ensure proper synchronization and prevent race conditions during cluster topology changes. [scylladb#27428](https://github.com/scylladb/scylladb/pull/27428) ### **Service Levels** - Ensured the service level cache is correctly updated after upgrading to v2 and prevented repeated service level upgrades when the cluster was already upgraded. [scylladb#28075](https://github.com/scylladb/scylladb/pull/28075) ### **Storage & SSTables** - Prevented “oversized allocation” errors when parsing sstable summary positions. [scylladb#24428](https://github.com/scylladb/scylladb/issues/24428), [scylladb#27590](https://github.com/scylladb/scylladb/issues/27590), [scylladb#27638](https://github.com/scylladb/scylladb/pull/27638) - Fixed an issue where Scylla failed to discover staging sstables on restart, leading to indefinite materialized view base-view inconsistency. [scylladb#27956](https://github.com/scylladb/scylladb/issues/27956), [scylladb#28090](https://github.com/scylladb/scylladb/pull/28090) - Ensured **truncate\_table\_on\_all\_shards** considers the **can\_flush** status on all shards. [scylladb#27639](https://github.com/scylladb/scylladb/issues/27639), [scylladb#28074](https://github.com/scylladb/scylladb/pull/28074) ### **Networking** - Resolved an issue where a recursive call in **posix\_server\_socket\_impl::accept()** could cause a stack overflow, ensuring greater stability in network connection handling. [scylladb#28166](https://github.com/scylladb/scylladb/issues/28166) ### **Performance** - Mitigated the **lsa - Aborting due to allocation failure** error that occurred during tablet population by implementing logic to dynamically allocate tablets based on the system’s current load. This prevents potential resource over-allocation during peak times, thereby improving overall stability and preventing critical allocation failures. [scylladb#27620](https://github.com/scylladb/scylladb/issues/27620), [scylladb#28060](https://github.com/scylladb/scylladb/pull/28060) - Added protection against negative count resource leaks in **reader\_concurrency\_semaphore** , enhancing the reliability of concurrency control mechanisms. [scylladb#28004](https://github.com/scylladb/scylladb/pull/28004) - Fixed a stall issue in system keyspace history retrieval for large numbers of tablets by ensuring a gentler synchronous unfreeze. [scylladb#27872](https://github.com/scylladb/scylladb/issues/27872), [scylladb#27909](https://github.com/scylladb/scylladb/pull/27909)