[prev in list] [next in list] [prev in thread] [next in thread] 

List:       forgerock-opendj
Subject:    [Opendj] OpenDJ 2.4.2 Reindex fails, takes userRoot offline
From:       matthew.swift () forgerock ! com (Matthew Swift)
Date:       2011-05-29 11:02:28
Message-ID: BANLkTimiYcZv1VffS7yAhwtnDDqV+wrwFg () mail ! gmail ! com
[Download RAW message or body]

I agree, this looks like a simple validation bug resulting in a division by
zero error to me which we should fix.

Matt


On Sun, May 29, 2011 at 7:30 AM, Mark Craig <mark.craig at gmail.com> wrote:

> Actually, Matt figures we shouldn't have to take the backend offline. See
> https://bugster.forgerock.org/jira/browse/OPENDJ-52
> 
> I still believe the behavior you found, Subu, is a bug. rebuild-index
> should perhaps fail with a message that the index does not exist instead.
> 
> Regards,
> Mark
> 
> 
> On Sun, May 29, 2011 at 7:21 AM, Mark Craig <mark.craig at gmail.com> wrote:
> 
> > Hello Subu,
> > 
> > The backend for the suffix does have to be taken offline while OpenDJ
> > rebuilds an index. If the backend could remain online while the index is
> > being rebuilt, then OpenDJ would have to be able to determine how to serve
> > operations that are supposed to be indexed but temporarily are not because
> > of the maintenance operation. No doubt that could be done, but the cost of
> > developing and testing the feature is probably not worth it. During the
> > rebuild, applications can in theory simply access another replica.
> > 
> > Yet IMHO the backend should not stay offline forever, even if you
> > mistakenly try to rebuild-index on an index that has not been created.
> > 
> > I see the same behavior as you do. (Pardon the French locale, I'm on the
> > kids' PC.)
> > 
> > PS C:\Users\Mark\OpenDJ\bat> ./rebuild-index -D "cn=directory manager" -p
> > 4444 -w password -X -b "dc=example,dc=com" -i audio
> > La t?che de Reconstituer l'index 20110529070931044 est planifi?e pour
> > d?marrer imm?diatement
> > [29/mai/2011:07:09:30 +0200] severity="NOTICE" msgCount=0 msgID=9896349
> > message="T?che Reconstituer l'index 20110529070931044 a commenc?
> > l'ex?cution"
> > [29/mai/2011:07:09:30 +0200] severity="INFORMATION" msgCount=1
> > msgID=9437595 message="Le backend de DB local userRoot n'indique pas le
> > nombre de threads de nettoyage?: il est par d?faut ? 97"
> > [29/mai/2011:07:09:30 +0200] severity="INFORMATION" msgCount=2
> > msgID=9437594 message="Le backend de DB local userRoot n'indique pas le
> > nombre de threads de nettoyage?: il est par d?faut ? 24 threads"
> > [29/mai/2011:07:09:31 +0200] severity="NOTICE" msgCount=3 msgID=9896306
> > message="Le backend userRoot est maintenant hors ligne "
> > [29/mai/2011:07:09:31 +0200] severity="SEVERE_ERROR" msgCount=4
> > msgID=10748756 message="Une erreur s'est produite lors de la tentative de
> > r?alisation de la reconstruction d'index?: / by zero"
> > [29/mai/2011:07:09:31 +0200] severity="NOTICE" msgCount=5 msgID=9896350
> > message="Reconstituer l'index t?che 20110529070931044 a termin? l'ex?cution"
> > La Reconstituer l'index t?che 20110529070931044 n'a pas ?t? termin?e
> > avec succ?s
> > 
> > After the operation completes unsuccessfully, dc=example,dc=com remains
> > unavailable, even if OpenDJ is restarted.
> > 
> > Would you mind logging a bug for this issue?
> > 
> > Regards,
> > Mark
> > 
> > On Sat, May 28, 2011 at 9:04 PM, Subu Ayyagari <subu.ayyagari at gmail.com>wrote:
> > 
> > > 
> > > Was a human error ... attempting a rebuild index without creating index
> > > in first place,
> > > took the suffix offline.
> > > 
> > > All ok now.
> > > 
> > > 
> > > On Sat, May 28, 2011 at 7:59 AM, Subu Ayyagari <subu.ayyagari at \
> > > gmail.com>wrote: 
> > > > Hi,
> > > > 
> > > > I tried to rebuild index for a few new attributes and it fails. It takes
> > > > suffix offline.
> > > > 
> > > > Is this a known issue/bug ? How do I recover ?
> > > > 
> > > > Thanks
> > > > 
> > > > 
> > > > *$ rebuild-index -X -b "dc=example,dc=com" -i automountkey -i
> > > > automountinformation *
> > > > 
> > > > 
> > > > 
> > > > [28/May/2011:07:25:10 -0400] severity="NOTICE" msgCount=3 msgID=9896306
> > > > message="The backend userRoot is now taken offline"
> > > > 
> > > > [28/May/2011:07:25:10 -0400] severity="SEVERE_ERROR" msgCount=4
> > > > msgID=10748756 message="An error occurred while attempting to perform
> > > > index rebuild:  / by zero"
> > > > 
> > > > [28/May/2011:07:25:13 -0400] severity="NOTICE" msgCount=5 msgID=9896350
> > > > message="Rebuild Index task 20110528072510681 finished execution"
> > > > 
> > > > 
> > > > 
> > > > *$ start-ds*
> > > > 
> > > > 
> > > > 
> > > > [28/May/2011:07:30:40 -0400] category=CONFIG severity=INFORMATION
> > > > msgID=3145839 msg=The backend defined in configuration entry \
> > > > ds-cfg-backend-id=userRoot,cn=Backends,cn=config is marked as disabled and \
> > > > therefore will not be used 
> > > > 
> > > > 
> > > > 
> > > 
> > > _______________________________________________
> > > OpenDJ mailing list
> > > OpenDJ at forgerock.org
> > > https://lists.forgerock.org/mailman/listinfo/opendj
> > > 
> > > 
> > 
> > 
> 
> _______________________________________________
> OpenDJ mailing list
> OpenDJ at forgerock.org
> https://lists.forgerock.org/mailman/listinfo/opendj
> 
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.forgerock.org/pipermail/opendj/attachments/20110529/7db279a8/attachment.html \



[prev in list] [next in list] [prev in thread] [next in thread]