[OpenAFS] Volumes going offline, needing salvage?

[Joseph Di Lellio]

A point of clarification: the old DB servers are not in the cell.
AFS has been turned off (verified n times), binaries/paths moved
or deleted, rc files gone, etc.  I have left them up for reference
with nothing running on them.

On Fri, 13 Oct 2006, Joseph Di Lellio wrote:

>
> Yes, I hit send rather earlier than intended.  Apologies.
>
> OpenAFS (v3.6): Sunfire V240 running Solaris 9 (118558-10)
> TransArc (v1.4.1): V220R running Solaris 8 (108528-09)
>
> There is nothing unusual in the system log files.  This is the
> case for all of the above, and the decom'ed DB servers.
>
> When I say unusual, I mean I still find the usual messages from
> reboots & such.
>
> The new fileservers are using a StorEdge 3510.  It has been configured
> as one 10 drive RAID 5 LUN which is divided into 3 equal partitions.
> There are 2 global spare drives.
>
> The old servers are using A1000s.
>
> The version of OpenAFS is the same across all DB & fileservers, and
> ditto for the TransArc systems.
>
> Filesystem type for /vicep* partitions are afs, with nologging.  There
> does not seem to be a correlation between the number of volumes & which
> of the two partitions I have.
>
> The entry:
>
> afs                     65
>
> exists in the /etc/name_to_sysnum on all systems.
>
> The volumes affected thus far are RW & BK.  I have seen one situation
> where both of these have had problems.  Others have had one or the other.
>
> I have a fairly limited number of RO volumes, so this may simply be
> statistics.
>
> The example below is a RW volume.  I have not seen differences between
> the volume types, including the combined instance above.
>
> An example of an entry from FileLog:
> Fri Oct 13 16:11:30 2006 VAttachVolume: volume salvage flag is ON for
> /vicepa//V0536976316.vol; volume needs salvage
>
> And from VolserLog:
> Fri Oct 13 16:11:30 2006 VAttachVolume: volume salvage flag is ON for
> /vicepa/V0536976316.vol; volume needs salvage
> Fri Oct 13 16:11:30 2006 1 Volser: XListOneVolume: Could not attach volume
> 536976316
>
> Looking back to the previous evening, I see a lot of the following:
>
> Thu Oct 12 23:57:20 2006 VAttachVolume: Error reading diskDataHandle vol
> header /vicepb/V0537103682.vol; error=101
> Thu Oct 12 23:57:20 2006 VAttachVolume: Error attaching volume
> /vicepb/V0537103682.vol; volume needs salvage; error=101
>
> Thu Oct 12 23:58:30 2006 VAttachVolume: Error reading smallVnode vol
> header /vicepb/V0536957965.vol; error=101
> Thu Oct 12 23:58:30 2006 VAttachVolume: Error attaching volume
> /vicepb/V0536957965.vol; volume needs salvage; error=101
>
> >From BosLog and the previous night, a number of these:
> Fri Oct 13 01:45:41 2006: fs:vol exited on signal 6 (core dumped)
>
> Although I see nothing similar on the other fileservers, they are
> otherwise identical.
>
> Apologies again for not including this earlier.  Your assistance is
> greatly appreciated.
>
> If there is anything else I'm missing, please let me know & I'll find
> it as quickly as I can.
>
> On Fri, 13 Oct 2006, Jeffrey Hutzelman wrote:
>
> >
> >
> > On Friday, October 13, 2006 04:32:45 PM -0700 Joseph Di Lellio
> > <joed@ucsc.edu> wrote:
> >
> > >    Does anyone have any ideas on what the issue(s) might be?  The logs
> > > have given me some bits, but mostly the obvious ones like needing to
> > > run salvage.
> >
> > You're going to need to provide some more details.
> > What operating system and version are the fileservers?
> > What OpenAFS version?
> > What filesystem type are the vice partitions?
> > Are the affected volumes RW, RO, or BK?
> > What are the relevant log messages?
> >
> > -- Jeff
> > _______________________________________________
> > OpenAFS-info mailing list
> > OpenAFS-info@openafs.org
> > https://lists.openafs.org/mailman/listinfo/openafs-info
> >
>
> ------
> It ain't what you don't know that gets you into trouble.  It's what you
> know for sure that just ain't so.		-- Mark Twain
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
>

------
It ain't what you don't know that gets you into trouble.  It's what you
know for sure that just ain't so.		-- Mark Twain