Benedict Elliott Smith created CASSANDRA-20325:
--------------------------------------------------
Summary: Accord: Fix recovery when one shard is committed/stable
and another is preaccepted
Key: CASSANDRA-20325
URL: https://issues.apache.org/jira/browse/CASSANDRA-20325
Project: Apache Cassandra
Issue Type: Bug
Components: Accord
Reporter: Benedict Elliott Smith
If one shard is committed/stable, but another is only preaccepted/accepted, we
must propose+stabilise on the remaining shards to ensure we create a correct
edge with recovery of earlier transactions on those shards
Also fix:
- CFK.visit removes transitive dependencies too eagerly across epoch change
- Topology slicing must declare whether we share/slice node ownership (to
assist above)
- apply cleanup to builder consistently, and construct the same value we
would produce by purge (so that replay is idempotent)
- Invoke ExecuteTxn.LocalExecute callbacks on originating CommandStore,
and add slow timeout once committed to help ensure progress if local store
cannot serve the read
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
