[
https://issues.apache.org/jira/browse/CASSANDRA-12954?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15719328#comment-15719328
]
Jeff Jirsa commented on CASSANDRA-12954:
----------------------------------------
Adding to all tables in {{system_distributed}} - we may not automatically
deleted from repair_history or parent_repair_history, but they're replicated,
so if a user issues deletes, it should get a proper gcgs to be safe.
CI pending, should be ready tomorrow:
|[3.X|https://github.com/jeffjirsa/cassandra/tree/cassandra-12954-3.X] |
[testall|http://cassci.datastax.com/job/jeffjirsa-cassandra-12954-3.X-testall/]
| [dtest|http://cassci.datastax.com/job/jeffjirsa-cassandra-12954-3.X-dtest/]|
|[trunk|https://github.com/jeffjirsa/cassandra/tree/cassandra-12954] |
[testall|http://cassci.datastax.com/job/jeffjirsa-cassandra-12954-testall/] |
[dtest|http://cassci.datastax.com/job/jeffjirsa-cassandra-12954-dtest/]|
> system_distributed.view_build_status uses gc_gs of 0, but issues distributed
> deletes
> ------------------------------------------------------------------------------------
>
> Key: CASSANDRA-12954
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12954
> Project: Cassandra
> Issue Type: Bug
> Components: Distributed Metadata
> Reporter: Aleksey Yeschenko
> Assignee: Jeff Jirsa
> Priority: Trivial
> Fix For: 3.x
>
>
> The definition uses {{CFMetaData.compile()}} method, intended for
> non-replicated system-keyspace tables, and doesn't override gc_grace_seconds
> from 0, but does issue deletes. This can lead to entries in the table not
> being cleaned up.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
