[OpenAFS] vos move & vos rel failing on large volumes
Todd DeSantis
atd@us.ibm.com
Thu, 10 Apr 2003 09:28:44 -0400
Hi -
Many times when I troubleshoot long running vos transactions,
I want to look at the traffic between the volservers.
I use the Meltdown.pl script that is available from the
Transarc IBM Web Site. It is just a perl script that is
a wrapper around the
rxdebug <src-volserver> 7003 -rxstat -noconn
Meltdown.pl -s <src volserver> -p 7005 -t 30 (seconds)
and I want to pay attention to the "resend" column.
If this column is increasing while the release/move is in
progress than I would want to look at the routes between
these 2 volserver machines.
Is duplex settings set properly on the network card and
switch port ? "netstat -ni" can show collisions if this is
a problem.
Can the network handle fragmented packets ?
How FULL is the destination vice partition ? Does it have
ample free inodes around ?
I would also look at the the volserver options
- using -udpsize 1048576
Are there any other vos tranactions happening on the src or
destination server during these times
vos status <src volserver>
vos status <dst volserver>
I would look into these things.
Thanks
Todd DeSantis
AFS Support
Nathan Neulinger
<nneul@umr.edu> To: Lee Damon <nomad@crow.ee.washington.edu>
Sent by: cc: openafs-info@openafs.org, nikola@crow.ee.washington.edu
openafs-info-admin@ Subject: Re: [OpenAFS] vos move & vos rel failing on large volumes
openafs.org
04/09/03 02:57 PM
That isn't a problem with the volume size... at least not _just_ the
volume size. We have numerous volumes > 10 GB with no problems at all on
linux, and move them around at will... Don't have many that large that a
replicated, but we definately have a few 2GB+ volumes that are
replicated.
-- Nathan
On Wed, 2003-04-09 at 13:01, Lee Damon wrote:
> Ignorant but trying to learn AFS...
>
> I know we shouldn't have volumes this large, but there isn't anything I
can
> do to fix the cause of them.
>
> fs exa
/afs/.ee.washington.edu/nikola/groups/vlsi/.Solaris8-sparc/pkgs/cadence/
> .2002a
> Volume status for vid = 536872898 named v.s8.cadence.2002a
> Current disk quota is 5000000
> Current blocks used are 2527255
> The partition has 12956012 blocks available out of 33223516
>
> However, I would still like to vos move and vos release them. Problem
is,
> vos move ends in the following failure. vos rel gives a similar error
> after which vos exa shows some volumes as new and some as old. Does
anyone
> have any hints about what I can check/change/do to try to get this move
to
> work?
>
> time vos move -fromserver kite -frompartition /vicepc -toserver potential
> -topartition /vicepf -id v.s8.cadence.2002a
> exit
> Failed to move data for the volume 536872898
> rxk: sealed data inconsistent
> vos move: operation interrupted, cleanup in progress...
> clear transaction contexts
> FATAL: VLDB access error: abort cleanup
> cleanup complete - user verify desired result
>
> real 619m16.383s
> user 0m0.004s
> sys 0m0.010s
>
> vos exa v.s8.cadence.2002a
> v.s8.cadence.2002a 536872898 RW 2527255 K On-line
> kite.ee.washington.edu /vicepc
> RWrite 536872898 ROnly 536873557 Backup 0
> MaxQuota 5000000 K
> Creation Wed Nov 20 15:54:32 2002
> Last Update Wed Mar 19 09:35:35 2003
> 0 accesses in the past day (i.e., vnode references)
>
> RWrite: 536872898
> number of sites -> 1
> server kite.ee.washington.edu partition /vicepc RW Site
>
>
> I'm getting errors from vos syncvldb now, as well:
>
> vos syncvldb potential /vicepf
> Could not create a VLDB entry for the volume 536873547
> VLDB: volume entry exists in the vldb
> Could not create a VLDB entry for the volume 536873457
> VLDB: volume entry exists in the vldb
> *** Warning: Orphaned RW volume 536872898 exists on
> potential.ee.washington.edu /vicepf
> VLDB reports RW volume 536872898 exists on kite.ee.washington.edu
/vicepc
> *** Warning: Orphaned RW volume 536872892 exists on
> potential.ee.washington.edu /vicepf
> VLDB reports RW volume 536872892 exists on kite.ee.washington.edu
/vicepc
> Could not process entries on server potential.ee.washington.edu partition
> /vicepf
> VLDB synchronized with state of server potential partition /vicepf
>
>
>
> Thanks for any advice,
> nomad
> ----------- - Lee "nomad" Damon - \
> work: nomad@ee.washington.edu \
> play: nomad@castle.org or castle!nomad \
> /\
> Sr. Systems Admin, UWEE SSLI Lab / \
> "Celebrate Diversity" / \
>
>
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
--
------------------------------------------------------------
Nathan Neulinger EMail: nneul@umr.edu
University of Missouri - Rolla Phone: (573) 341-4841
Computing Services Fax: (573) 341-4216
_______________________________________________
OpenAFS-info mailing list
OpenAFS-info@openafs.org
https://lists.openafs.org/mailman/listinfo/openafs-info