[Linux-cluster] Qdiskd not working anymore
billpp at gmail.com
Sat Dec 13 06:06:24 UTC 2008
-----BEGIN PGP SIGNED MESSAGE-----
When you have more than one path to access a storage-device (SAN), you
will need to install/configure some kind of MultiPath I/O.
IBM has RDAC, EMC has PowerPath, HP has SecurePath, and so on..
Linux device-mapper (md) has dm-multipath as an alternative for
proprietary drivers above cited, maybe you can use it to configure
your multipath storage.
BTW, this is off-topic here (if your problem really is multipath
configuration, as i guessed).
Flávio do Carmo Júnior aka waKKu
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
-----END PGP SIGNATURE-----
On Sat, Dec 13, 2008 at 3:57 AM, Maykel Moya <moya-lists at infomed.sld.cu>wrote:
> I have a SAN LUN exported to a four node cluster. I' m using Quorum Disk
> and it was working OK (ie: shows up in clustat).
> The cluster isn't yet in production. Some days ago another io path was
> added to the SAN.
> Today I issued a clustat and see no reference to qdiskd. The daemon
> refused to start:
> e1b01:~# qdiskd -fd
>  debug: Loading configuration information
>  debug: 0 heuristics loaded
>  debug: Quorum Daemon: 0 heuristics, 3 interval, 23 tko, 3 votes
>  debug: Run Flags: 00000035
>  crit: Specified device /dev/sda1 does match kernel's reported
> sector size (0 != -1)
> I wonder is there are any relationship between qdiskd failing and seeing
> the LUN partition twice in each node. /dev/sda1 is the qdisk, also seen
> as /dev/sdb1.
> I tried to recreate the quorum disk and updated cluster.conf to use
> device instead of label but still qdiskd refuse to start
> I'm using:
> linux 2.6.27
> rhcs 2.03.08
> openais 0.83
> on debian lenny.
> Thanks very much for your help
> Red Telematica de Salud - Cuba
> CNICM - Infomed
> Linux-cluster mailing list
> Linux-cluster at redhat.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Linux-cluster