You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@bookkeeper.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2017/06/28 21:32:00 UTC

[jira] [Commented] (BOOKKEEPER-1098) ZkUnderreplicationManager can build up an unbounded number of watchers

    [ https://issues.apache.org/jira/browse/BOOKKEEPER-1098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16067267#comment-16067267 ] 

ASF GitHub Bot commented on BOOKKEEPER-1098:
--------------------------------------------

Github user athanatos commented on the issue:

    https://github.com/apache/bookkeeper/pull/193
  
    @sijie Oops, I just noticed the comments.  I guess you just changed the title?


> ZkUnderreplicationManager can build up an unbounded number of watchers
> ----------------------------------------------------------------------
>
>                 Key: BOOKKEEPER-1098
>                 URL: https://issues.apache.org/jira/browse/BOOKKEEPER-1098
>             Project: Bookkeeper
>          Issue Type: Bug
>          Components: bookkeeper-auto-recovery
>            Reporter: Samuel Just
>            Assignee: Samuel Just
>            Priority: Minor
>             Fix For: 4.5.0
>
>
> getLedgerToReplicate leaves watches each time it traverses the
> tree until it finds a suitable replication target.  Since we don't have
> a way of canceling watches, these watches tend to get abandoned,
> particularly on interior nodes, which aren't changed much.  Thus,
> over time, some nodes would build up a very large number of watches.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)