The ScyllaDB team announces the release of ScyllaDB Open Source 3.1.2, a bugfix release of the ScyllaDB 3.1 stable branch. ScyllaDB Open Source 3.1.2, like all past and future 3.x.y releases, is backward compatible and supports rolling upgrades.
Note: if and only if you installed a fresh ScyllaDB 3.1.0, you must add the following line to scylla.yaml of each node before upgrading to 3.1.2:
enable_3_1_0_compatibility_mode: true
This is not relevant if your cluster was upgraded to 3.1.0 from an older version, or you are upgrading from or to any other ScyllaDB releases, like 3.1.1.
If you have doubts, please contact us using the user mailing list.
Related links:
- ScyllaDB Open Source 3.1
- Get ScyllaDB Open Source 3.1.2 – Docker, binary packages, and EC2 AMI
- Get started with ScyllaDB
- Upgrade from ScyllaDB Open Source 3.x.y to ScyllaDB Open Source 3.x.z
- Upgrade from ScyllaDB Open Source 2.3.x to 3.0.x
- Please let us know if you encounter any problems.
Issues fixed in this release
- CQL: wrong key type used when creating non-frozen map virtual column #5165
- CQL: One second before expiration, TTLed columns return as null values #4263, #5290
- CQL: using queries with paging, ALLOW FILTERING and aggregation functions return intermediate aggregated results, not the full one #4540
- CQL Local Index: potentially incorrect partition slices might cause a minor performance impact #5241
- Stability: non-graceful handling of end-of-disk space state, may cause ScyllaDB to exit with a coredump #4877
- Stability: core dump on OOM during cache update after memtable flush, with
!_snapshot->is_locked()' failederror message #5327 - Stability: long-running cluster sees bad gossip generation when a node restarts #5164 (similar to CASSANDRA-10969)
- Stability: Oversized allocation warning in reconcilable_result, for example, when paging is disabled #4780
- Stability: running manual operations like nodetool compact will crash if the controller is disabled #5016
- Stability: possible use-after-free during shutdown #5242
- Stability: Under heavy read load, the read execution stage queue size can grow without bounds #4749
- Stability: a rare race condition can cause a crash when updating schema during stress view updates #4856, #4766
- Stability: certain node failures during repair are not handled correctly, might cause ScyllaDB to exit #5238
- UX: sstables:
delete_atomically: misplaced parenthesis in warning message #4861 - UX: scylla_setup script: NIC selection prompt causes error when Enter is pressed key without a NIC name #4517
- Build from source:
libthread_db.so.1is missing in the relocatable package installation causing GDB thread debugging to be disabled #4996 - Build from source: relocatable builds (via
reloc/build_reloc.sh) are not reproducible #5222
03 Dec 2019
