[Linux-cluster] quotaoff -gu causes problems

[Pavlos Parissis]

Quoting Pavlos Parissis <p_pavlos at freemail.gr>:

> Hi,
>
> I have a service which mounts 3 ext3 FS and then share it over a nfs -
> this is done manually by a script because nfs exports "are" broken in
> 4.u4.
> For more info about this read
> https://www.redhat.com/archives/linux-cluster/2008-February/msg00184.html.
>
> When the exported file systems are mounted by the clients with options
> bg,rw,rsize=8192,wsize=8192,intr,hard and try to fail over the
> service, the stopping process hangs when it executes quotaoff -gu.
> I see that in ps -ef and when I run manually quotaoff -gu hangs as well.
>
> I thought to put noquota on the options for the fs but it didn't help
> because the /usr/share/cluster/fs.sh runs quotaoff -gu regardless if
> the fs are mounted with noquota.
>
> Any idea how to solve this?
>
> Regards,
> Pavlos
>

here is my cluster conf for the service
question about the fsid - do I put any  number I want ?

</service>
     <service domain="ThirdDomain" name="nfs1" autostart="1"  
recovery="relocate">
       <fs device="/dev/sdb2" options="noquota" nfslock="1"  
fsid="1002" mountpoint="/nfs_omg_data" force_unmount="1"  
name="/nfs_omg_data" fstype="ext3"/>
       <fs device="/dev/sdb3" options="noquota" nfslock="1"  
fsid="1003" mountpoint="/nfs_apps" force_unmount="1" name="/nfs_apps"  
fstype="ext3"/>
       <fs device="/dev/sdb4" options="noquota" nfslock="1"  
fsid="1004" mountpoint="/nfs_cdr_central" force_unmount="1"  
name="/nfs_cdr_central" fstype="ext3"/>
       <ip monitor_link="1" address="10.10.21.73"/>
       <ip monitor_link="1" address="10.10.21.142"/>
       <script name="nfs1" file="/usr/local/wsb/scripts/rhc_nfs1"/>
     </service>