[torqueusers] Newbie questions
knielson at adaptivecomputing.com
Fri Mar 21 10:59:23 MDT 2014
On Fri, Mar 21, 2014 at 10:55 AM, David Beer <dbeer at adaptivecomputing.com>wrote:
> On Fri, Mar 21, 2014 at 5:57 AM, Dr.Peer-Joachim Koch <
> pkoch at bgc-jena.mpg.de> wrote:
>> I just started to play around a bit with torque.
>> The installation worked without any problems (torque-4.2.7).
>> For the configuration I still have some questions:
>> *) we have a shared file system. So every node sees the same
>> file system layout for groups, home and /usr/local.
>> How do I have to configure it in mom_priv/config ?
> I hope someone else can also offer an opinion here, but often times people
> put mom_priv and mom_log as separate directories for each mom, but the
> config file is a symlink to somewhere on the shared file system.
>> /User/homes/USER ist the path to the users $HOME.
>> Also stored in LDAP (homeDirectory)
>> $usecp *:/User/homes /User/homes ?
> This (without the question mark) looks to me to say that any directory
> that is /User/homes on any machine is viewable on any of the compute nodes.
>> *) How can I add the domain for sending e-mail ?
> I believe mail_domain is the parameter you're looking for:
> qmgr -c 'set server mail_domain=<desired domain>'
>> *) I asked before, if I can install torque (for testing) while LSF is
>> still active (and used inthe moment). It worked so far,
>> but 2h after I started the master node, this node crashed (reboot)
>> without any usefull log information. Might it interact in some strange
>> way ? i've tested it again (no time), so I want to ask before I crash
>> some nodes of our cluster ... ;)
> I would recommend avoiding this setup if possible. At the least, no one
> tests that our software is able to do this and it seems likely that they
> would interfere with each others' jobs and general operating. If possible,
> I'd recommend selecting a few nodes from the cluster to remove from LSF's
> view, configure for TORQUE, and test. I think that's most likely to give
> you a picture of how TORQUE behaves.
One more note about LSF and TORQUE. TORQUE is just a resource manager. It
needs a scheduler to automatically run jobs. What are you using as the
>> Thanks and bye,
>> Max-Planck-Institut fuer Biogeochemie
>> Dr. Peer-Joachim Koch
>> Hans-Knöll Str.10 Telefon: ++49 3641 57-6705
>> D-07745 Jena Telefax: ++49 3641 57-7705
>> torqueusers mailing list
>> torqueusers at supercluster.org
> David Beer | Senior Software Engineer
> Adaptive Computing
> torqueusers mailing list
> torqueusers at supercluster.org
[image: Adaptive Computing] <http://www.adaptivecomputing.com>
[image: Twitter] <http://twitter.com/AdaptiveMoab> [image: LinkedIn]
YouTube] <http://www.youtube.com/adaptivecomputing> [image: GooglePlus]
<https://plus.google.com/u/0/102155039310685515037/posts> [image: Facebook]
Ken Nielson Sr. Software Engineer
+1 801.717.3700 office +1 801.717.3738 fax
1712 S. East Bay Blvd, Suite 300 Provo, UT 84606
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the torqueusers