www-infrastructure-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "#asfinfra IRC Bot (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (INFRA-4685) Open file handles problem on lucene.zones.apache.org
Date Thu, 19 Apr 2012 10:52:42 GMT

    [ https://issues.apache.org/jira/browse/INFRA-4685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13257422#comment-13257422
] 

#asfinfra IRC Bot commented on INFRA-4685:
------------------------------------------

<danielsh> The offending jail's admins are looking into the problem and have taken steps
to prevent its recurrence.

                
> Open file handles problem on lucene.zones.apache.org
> ----------------------------------------------------
>
>                 Key: INFRA-4685
>                 URL: https://issues.apache.org/jira/browse/INFRA-4685
>             Project: Infrastructure
>          Issue Type: Task
>      Security Level: public(Regular issues) 
>          Components: Zones/Jails
>            Reporter: Dawid Weiss
>
> Hi. We've been running into errors caused by exceeding the number of file handles recently
on jenkins (lucene.zones.apache.org). From what we can tell:
> {noformat}
> kern.maxfiles: 12328
> kern.maxfilesperproc: 11095
> kern.openfiles: 11601
> {noformat}
> so it's indeed very close to the system maxfiles. We have tried to pinpoint the problem
but we failed (Robert Muir, Dawid Weiss). Our suspicion is that something else (a different
jail?) is sucking all the resources out but it's hard to tell, really.
> Quesitons/ wishes:
> - could you take a look at the machine's jails -- is maxfiles shared between them or
is it per-jail limit? If it's per-jail then what is eating away so many file handles on lucene.*?
> - would it be possible to bump maxfiles (if the cause of the problem remains unknown)
so that we can proceed with regular builds?
> Any other suggestion you may provide to resolve the problem will be very welcome. We
lsof'ed, sockstated and everything else. Solr/Lucene builds are pretty heavy but the above
situation is present even when no build takes place.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message