[
https://issues.apache.org/jira/browse/CASSANDRA-15690?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sam Tunnicliffe updated CASSANDRA-15690:
----------------------------------------
Test and Documentation Plan: New in-jvm dtests added. For trunk, the
proposed snapshot on mismatch feature has a descriptive comment in
cassandra.yaml, but we should add full docs for repaired data tracking in
general.
Status: Patch Available (was: Open)
Pushed branches with fixes and additional tests for 3.0, 3.11 and trunk. The
trunk branch includes a second commit which adds the ability to trigger a
snapshot if a mismatch is detected between repaired data across replicas either
at query time (using the tracking capability from CASSANDRA-14145) or during a
preview repair. The snapshotting is controlled by a yaml setting and can also
be enabled/disabled by jmx. To avoid filling the disks with snapshots, each
replica will take at most 1 snapshot per-table per-day via this mechanism
(manually triggered snapshots are not affected). These snapshots can be very
useful in debugging temporary divergences between replicas which may be
resolved by full or read repairs before investigation takes place.
||branch||utests||in-jvm dtests||dtests_with_vnodes||dtests_no_vnodes||
|[15690-3.0|https://github.com/beobal/cassandra/tree/15690-3.0]|[jdk8|https://circleci.com/gh/beobal/cassandra/1290]|[jdk8|https://circleci.com/gh/beobal/cassandra/1289]|[jdk8|https://circleci.com/gh/beobal/cassandra/1296]|[jdk8|https://circleci.com/gh/beobal/cassandra/1297]|
|[15690-3.11|https://github.com/beobal/cassandra/tree/15690-3.11]|[jdk8|https://circleci.com/gh/beobal/cassandra/1311]|[jdk8|https://circleci.com/gh/beobal/cassandra/1312]|[jdk8|https://circleci.com/gh/beobal/cassandra/1314]|[jdk8|https://circleci.com/gh/beobal/cassandra/1313]|
|[15690-trunk|https://github.com/beobal/cassandra/tree/15690-trunk]|[jdk8|https://circleci.com/gh/beobal/cassandra/1303],
[jdk11|https://circleci.com/gh/beobal/cassandra/1307]|[jdk8|https://circleci.com/gh/beobal/cassandra/1304],
[jdk11|https://circleci.com/gh/beobal/cassandra/1302]|[jdk8|https://circleci.com/gh/beobal/cassandra/1305],
[jdk11|https://circleci.com/gh/beobal/cassandra/1309]|[jdk8|https://circleci.com/gh/beobal/cassandra/1306],
[jdk11|https://circleci.com/gh/beobal/cassandra/1308]|
> Single partition queries can mistakenly omit partition deletions and
> resurrect data
> -----------------------------------------------------------------------------------
>
> Key: CASSANDRA-15690
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15690
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Coordination
> Reporter: Aleksey Yeschenko
> Assignee: Sam Tunnicliffe
> Priority: Normal
> Fix For: 3.0.x, 3.11.x, 4.0-alpha
>
>
> We have logic that allows us to exclude sstables with partition deletions
> that are older than the minimum collected timestamp in a local request.
> However, it’s possible that another node could have rows that aren’t known to
> the local node that are in turn older than the excluded partition deletion.
> In such a scenario, those will be mistakenly resurrected, which is a
> correctness issue.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
