[Linux-cluster] error clusvcadm

Ryan Mitchell rmitchel at redhat.com
Tue May 14 00:42:32 UTC 2013


On 05/13/2013 06:48 PM, emmanuel segura wrote:
> Hello Delphine
>
> your problem as you know is here
> ===============================================================
> <fs device="LABEL=postfix" mountpoint="/var/spool/
> postfix" force_unmount="1" fstype="ext3" name="mgmtha5" options=""/>
> ===============================================================
>
> I don't know if you are using lvm or partition, but you should look for
> the device corresponding to that LABEL, if you are using lvm use vgs and
> lvs to see if your volume are actived
>
> Thanks

Further to this:
- It is best-practice to use LVM resources (HA-LVM) for non-clustered 
filesystems (ie. ext3, ext4, xfs, etc).  This helps to protect against 
accidentally mounting the filesystem on more than one node at a time.

As you have some GFS2, I assume you have clustered LVM and therefore it 
is as simple as adding LVM resources to the services.  See 
https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Cluster_Administration/ap-ha-halvm-CA.html#s1-halvm-clvm-CA

- The error you're experiencing where the service won't start is because 
the filesystem with the label "postfix" can't be found (as Emmanuel 
said).  See if you can see the missing device with the following command:
# blkid -c /dev/null

If there is no device with LABEL=postfix, you need to try to find which 
is the correct device, and add the label to the device with e2label.

Regards,

Ryan Mitchell
Red Hat Global Support Services




More information about the Linux-cluster mailing list