[Ocfs2-tools-users] clone production luns issue

[Sunil Mushran]

mount that device once. mount and then umount.
fsck replayes the journal but does not clear the
dirty journal bit. Hmmm... file a bugzilla. We need
to address this.

Antonio Marongiu wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi All,
>
> We are trying to clone production lun  to refresh Oracle preproduction environments and to backup production enviroments.
>
> This is my environment,
>
> 2 nodes 10GR2  RAC redhat AS4
>
> these are my production luns
>
> /dev/dm-1             1.0G  286M  739M  28% /u03    (cluster registry)
> /dev/dm-0             601G   26G  575G   5% /u04/siebit/datafile01
> /dev/dm-3              10G  3.5G  6.6G  35% /u05/siebit/oralogs01
> /dev/dm-2              50G   39G   12G  77% /u06/siebit/archive01
>
>
> I've installed the following ocfs2 packages
>
> ocfs2console-1.2.4-1
> ocfs2-2.6.9-42.0.3.ELsmp-1.2.5-2
> ocfs2-tools-1.2.4-1
>
>
>
> We are using netapp lun clone  and the cloning is working properly.  I trying  to mount up the newly cloned luns on the production environment where
> there are also mounted the production lun and have several problems.
>
>
> [root at node01 ~]# mount /backup/siebit/datafile01
> mount.ocfs2: Invalid argument while mounting /dev/dm-4 on /backup/siebit/datefile01. Check 'dmesg' for more information on this error.
>
>
> Checking dmesg I saw the following error:
>
> (24069,7):ocfs2_fill_super:620 ERROR: Unable to create per-mount debugfs root.
> ocfs2: Unmounting device (253,4) on (node 255)
>
>
> These error probably is caused by the cloned lun have the same UUID and conflicts with production lun.
>
>
> I tried to modify cloned lun uuid using "tunefs.ocfs2 --uuid-reset"  command but I can't succeed, I gave the following error:
>
>
> [root at siebel-db-it01 ~]# tunefs.ocfs2 -U /dev/dm-4
> tunefs.ocfs2 1.2.4
> tunefs.ocfs2: Trylock failed while locking down the cluster
>
>
> This error doesn't appear if I shutdown the RAC and unmount the all production luns, at this point I'm able to perform tunfs.ocsf2 on luns dm-4 dm-6,
> but  I gave the following error while performing tunefs.ocfs2 -U on the lun /dev/dm-5
>
> [root at node01 ~]# tunefs.ocfs2 -U /dev/dm-5
> tunefs.ocfs2 1.2.4
> tunefs.ocfs2: Node slot 0's journal is dirty. Run fsck.ocfs2 to replay all dirty journals.
>
>
> [root at node01 ~]# fsck.ocfs2 /dev/dm-5
> Checking OCFS2 filesystem in /dev/dm-5:
>   label:              oradata01
>   uuid:               31 c8 80 ab e8 18 49 02 b6 7b d0 d7 81 35 4b b8
>   number of blocks:   157306368
>   bytes per block:    4096
>   number of clusters: 1228956
>   bytes per cluster:  524288
>   max slots:          4
> /dev/dm-5 is clean.  It will be checked after 20 additional mounts.
>
> /dev/dm-5 seems ok, but tunefs.ocfs2 still not working
>
> [root at node01 ~]# tunefs.ocfs2 -U /dev/dm-5
> tunefs.ocfs2 1.2.4
> tunefs.ocfs2: Node slot 0's journal is dirty. Run fsck.ocfs2 to replay all dirty journals.
>
>
> Where's mistake? Does anyone have a procedure to perform a ocsf2 cloning?
>
>
> Thanks in advance,
>
>
>
> Antonio
>
>
>
>
>
>
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFGjl/eHevQJWvdvLYRAqpXAJ41PWZEs6NZJcejk2WyBkkLlbSutwCeK7Kz
> uwopMgoshiDGoTqyijVkQLE=
> =RW4D
> -----END PGP SIGNATURE-----
>
> _______________________________________________
> Ocfs2-tools-users mailing list
> Ocfs2-tools-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-tools-users
>