On 03/24/2010 08:18 PM, Alex Dean wrote:
> 
> On Mar 24, 2010, at 5:11 PM, Terry Inzauro wrote:
> 
>> On 03/24/2010 01:12 PM, Dimitri Maziuk wrote:
>>> On Tuesday 23 March 2010 18:39:40 Terry Inzauro wrote:
>>>
>>>> r...@valerie:~# mount | grep rpc
>>>> rpc_pipefs on /data/shared_nfs_data/rpc_pipefs type rpc_pipefs (rw)
>>>>
>>>> r...@valerie:~# ls -l /var/lib | grep nfs
>>>> lrwxrwxrwx  1 root      root        22 2010-03-23 08:49 nfs ->
>>>> /data/shared_nfs_data/
>>>>
>>>>
>>>> Do i need to manually change the mount point somewhere as to not
>>>> reference
>>>> the sym linked path?
>>>
>>> Yes. See www.linux-ha.org/HaNFS -- you need to reconfigure idmapd and
>>> the
>>> kernel module.
>>>
>>> Dima
>>
>> Will do. I'll read that doc and return my results. Thank you again for
>> the assist.
>>
> 
> FWIW, I've used the www.linux-ha.org/HaNFS instructions to set up
> several NFS-serving HA clusters using Red Hat 5, and it's worked exactly
> as advertised.
> 
> alex
> 
> 



Ok. I got it figured out now. Wow, that was fun.  Basically, I misread or 
misinterpreted the document.

I initially (and presumptuously) moved /var/lib/nfs to my shared storage and 
then sym linked it back to /var/lib/nfs. Which
means that rpc_pipefs would now me mounted on shared storage.  After changing 
the mount point for rpc_pipcs back to unshared
file system (/var/lib/rpc_pipefs) all worked brilliantly.

One thing to note:
On Debian Lenny, the mount point for rpc_pipefs needs to be changed in 
/etc/init.d/nfs-common and in /etc/idmapd.conf.  By
default, the mount point is hard coded in /etc/init.d/nfs-common and not 
defined in /etc/default/nfs-common.  I commented out
the mount point in the init script, and added the variable to 
/etc/default/nfs-common.  This was done so my changes don't get
clobbered upon upgrade.

One last item.  When I add resources with crm < /path/to/resources, they 
changes don't get saved to file system and
subsequently do not get loaded if all cluster nodes are rebooted.

Is this intended behaviour or have I missed something?


kind regards,


Terry



_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to