[torqueusers] Torque HA on Lustre

Michael Robbert mrobbert at mines.edu
Fri Sep 12 12:16:53 MDT 2008


I just wanted to let everybody know that we have Torque HA working on  
Lustre now. There was a lot of things going on around the same time,  
but I think that the fix for this problem was upgrading Lustre to  
1.6.5 from 1.6.4.3

Mike Robbert, CSM
On Jul 24, 2008, at 12:58 AM, Frank Mietke wrote:

> Hi Mike,
>
> On Wed, Jul 23, 2008 at 09:46:39AM -0600, Michael Robbert wrote:
>> Is anybody on the list doing Torque High Availability using Lustre  
>> as their
>> shared file system for server_priv? We are trying to do this, but can
>> Lustre keeps falling over on us. I know very little about Lustre,  
>> but our
>> vendors are having trouble debugging this and have recently  
>> speculated that
>> there may be a locking issue. We are currently testing with  
>> server_priv on
>
> I tried fully installing to and running torque from lustre. But  
> stopped due to
> lack of flock support. Now I have server_priv installed locally and  
> the rest in
> a lustre filesystem. Maybe you're able to disable the flock support  
> in torque. I
> have not tried this.
>
> In chapter 16 of Lustre manual you can find the following:
>
> Note  Advisory fcntl/flock/lockf locks will be available in Lustre  
> 1.8.
>
> Please, keep me informed if you have success.
>
> Best Regards,
> Frank
>
>
>> an NFS share, but that share is on one of the head nodes so that is  
>> not an
>> optimal HA solution. I'm just curios if anybody else is doing this
>> successfully and if so have you done anything special with Lustre  
>> or Torque
>> to make it work properly.
>>
>> Thanks,
>> Mike Robbert
>> _______________________________________________
>> torqueusers mailing list
>> torqueusers at supercluster.org
>> http://www.supercluster.org/mailman/listinfo/torqueusers
>>
>
> -- 
> Dipl.-Inf. Frank Mietke     |     Fakultätsrechen- und  
> Informationszentrum
> Tel.: 0371 - 531 - 35538    |     Fak. für Informatik
> Fax:  0371 - 531 8 35538    |     TU-Chemnitz
> Key-ID: 60F59599            |     frank.mietke at informatik.tu-chemnitz.de



More information about the torqueusers mailing list