[jira] [Commented] (CASSANDRA-12504) BatchlogManager is shut down twice during drain

["Stefania (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-12504?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15432211#comment-15432211 ] 

Stefania commented on CASSANDRA-12504:
--------------------------------------

I agree, the first call [here|https://github.com/stef1927/cassandra/commit/218206434d66cbe1d00191d2248474745d132eed#diff-b76a607445d53f18a98c9df14323c7ddR3951] looks like the correct place. 

[~iamaleksey], this looks like an oversight from CASSANDRA-9673, so I am assigning you as a reviewer in case your memory is better of mine and can find a reason why we wanted to move it to the end.

Here is the patch for 3.0, I'm not wasting cycles repeating CI on 3.9 and trunk for this unless you feel we should:

|[patch|https://github.com/stef1927/cassandra/commits/12504-3.0]|[testall|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-12504-3.0-testall/]|[dtest|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-12504-3.0-dtest/]|


> BatchlogManager is shut down twice during drain
> -----------------------------------------------
>
>                 Key: CASSANDRA-12504
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12504
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Alex Petrov
>            Assignee: Stefania
>            Priority: Minor
>             Fix For: 3.0.x, 3.x
>
>
> {{BatchlogManager}} is shut down twice during in the {{StorageService}}, once [here|https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/service/StorageService.java#L4216] and once [here|https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/service/StorageService.java#L4285].
> The first shutdown was already there but the last one seems to have been added [here|https://github.com/apache/cassandra/commit/53a177a9150586e56408f25c959f75110a2997e7#diff-b76a607445d53f18a98c9df14323c7ddR3913]. It seems to be harmless, as it’s not required in stages in-between, so second run would be a no-op.
> Following the logic of other shutdown hook, the first one is the good place for it to be (right before {{HintsService}}).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)