You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by "Eric Newton (JIRA)" <ji...@apache.org> on 2012/09/27 23:11:09 UTC
[jira] [Created] (ACCUMULO-779) ZooStore can starve completed Repos
Eric Newton created ACCUMULO-779:
------------------------------------
Summary: ZooStore can starve completed Repos
Key: ACCUMULO-779
URL: https://issues.apache.org/jira/browse/ACCUMULO-779
Project: Accumulo
Issue Type: Bug
Components: master
Affects Versions: 1.4.0, 1.4.1
Reporter: Eric Newton
Assignee: Eric Newton
Fix For: 1.4.2
A user aggressively used the compactrange command. This queued dozens of Repos in zookeeper. Pulling new Repos from zookeeper is fairly slow. Checking if they are ready, is also slow. The threads running the Repos would not finish getting through the list of Repos, some of which had finished, before the first item in the list was ready to be checked again.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (ACCUMULO-779) ZooStore can starve completed Repos
Posted by "Keith Turner (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Keith Turner updated ACCUMULO-779:
----------------------------------
Priority: Critical (was: Major)
> ZooStore can starve completed Repos
> -----------------------------------
>
> Key: ACCUMULO-779
> URL: https://issues.apache.org/jira/browse/ACCUMULO-779
> Project: Accumulo
> Issue Type: Bug
> Components: master
> Affects Versions: 1.4.1, 1.4.0
> Reporter: Eric Newton
> Assignee: Keith Turner
> Priority: Critical
> Fix For: 1.5.0, 1.4.2
>
>
> A user aggressively used the compactrange command. This queued dozens of Repos in zookeeper. Pulling new Repos from zookeeper is fairly slow. Checking if they are ready, is also slow. The threads running the Repos would not finish getting through the list of Repos, some of which had finished, before the first item in the list was ready to be checked again.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Assigned] (ACCUMULO-779) ZooStore can starve completed
Repos
Posted by "Keith Turner (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Keith Turner reassigned ACCUMULO-779:
-------------------------------------
Assignee: Keith Turner (was: Eric Newton)
> ZooStore can starve completed Repos
> -----------------------------------
>
> Key: ACCUMULO-779
> URL: https://issues.apache.org/jira/browse/ACCUMULO-779
> Project: Accumulo
> Issue Type: Bug
> Components: master
> Affects Versions: 1.4.1, 1.4.0
> Reporter: Eric Newton
> Assignee: Keith Turner
> Fix For: 1.5.0, 1.4.2
>
>
> A user aggressively used the compactrange command. This queued dozens of Repos in zookeeper. Pulling new Repos from zookeeper is fairly slow. Checking if they are ready, is also slow. The threads running the Repos would not finish getting through the list of Repos, some of which had finished, before the first item in the list was ready to be checked again.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (ACCUMULO-779) ZooStore can starve completed
Repos
Posted by "Keith Turner (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-779?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13465110#comment-13465110 ]
Keith Turner commented on ACCUMULO-779:
---------------------------------------
The txids are random. The repos that were basically finished were holding locks and had higher txids. Fate kept cycling through the lower txids, that were just waiting on the lock, never getting to the higher txids.
Need to make sure the reservation function cycles through all txids before starting over.
> ZooStore can starve completed Repos
> -----------------------------------
>
> Key: ACCUMULO-779
> URL: https://issues.apache.org/jira/browse/ACCUMULO-779
> Project: Accumulo
> Issue Type: Bug
> Components: master
> Affects Versions: 1.4.1, 1.4.0
> Reporter: Eric Newton
> Assignee: Eric Newton
> Fix For: 1.4.2
>
>
> A user aggressively used the compactrange command. This queued dozens of Repos in zookeeper. Pulling new Repos from zookeeper is fairly slow. Checking if they are ready, is also slow. The threads running the Repos would not finish getting through the list of Repos, some of which had finished, before the first item in the list was ready to be checked again.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (ACCUMULO-779) ZooStore can starve completed
Repos
Posted by "Keith Turner (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Keith Turner resolved ACCUMULO-779.
-----------------------------------
Resolution: Fixed
Fix Version/s: 1.5.0
> ZooStore can starve completed Repos
> -----------------------------------
>
> Key: ACCUMULO-779
> URL: https://issues.apache.org/jira/browse/ACCUMULO-779
> Project: Accumulo
> Issue Type: Bug
> Components: master
> Affects Versions: 1.4.1, 1.4.0
> Reporter: Eric Newton
> Assignee: Eric Newton
> Fix For: 1.5.0, 1.4.2
>
>
> A user aggressively used the compactrange command. This queued dozens of Repos in zookeeper. Pulling new Repos from zookeeper is fairly slow. Checking if they are ready, is also slow. The threads running the Repos would not finish getting through the list of Repos, some of which had finished, before the first item in the list was ready to be checked again.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira