The ScyllaDB team is pleased to announce the release of ScyllaDB 1.3.2, a bugfix release of the ScyllaDB 1.3 stable branch. Release 1.3.2, like all past and future 1.x.y releases, is backward compatible and supports rolling upgrades.
The following bugs have been fixed:
- Fix missing information in
nodetool cfstatsreport when LeveledCompactionStrategy is used #1119 - Fix incorrect handling of clustering row ranges in paging code, which resulted in not being able to select more than 100 rows when filtering with clustering key in cqls cqlsh #1684
- Fix gossip expiration timer inconsistency across a cluster by using system clock instead of steady clock #1704
- Fix CQL server request throttling so that requests don’t overflow memory and cause std::bad_alloc exceptions #1708
- Fix too early gossip failure_detector history removal, which can lead to a node thinking a decommissioned node is still alive #1714
- Fix sstable readers created for streaming from timing out too soon #1741
Related links
- ScyllaDB 1.3 release
- Get started with ScyllaDB 1.3.2
- Upgrade from 1.2.x to 1.3.y
- Report a problem
ScyllaDB contributions
Asias He (2):
gossip: Do not remove failure_detector history on remove_endpoint
gossip: Switch to use system_clock
Avi Kivity (1):
Update seastar submodule
Paweł Dziepak (1):
query_pagers: fix clustering key range calculation
Pekka Enberg (2):
release: prepare for 1.3.2
database: Fix io_priority_class related compilation error
Raphael S. Carvalho (1):
api: implement api to return sstable count per level
Tomasz Grabiec (2):
transport: Extend request memory footprint accounting to also cover execution
db: Do not timeout streaming readers
