[torqueusers] how can i change mom_log directory

Yanan Sun nancyprc at gmail.com
Thu Jul 31 09:09:06 MDT 2008


thanks, Glen
Kamil, what i did at the end is create a folder has all nodes
pbs_spool files under /mnt/PBS/

[root at master ~]# ls /mnt/PBS/
lost+found  node001  node002  node003

[root at master ~]# ls /mnt/PBS/node001/*
aux         mom_logs  pbs_environment  sched_priv   server_name  spool
checkpoint  mom_priv  sched_logs       server_logs  server_priv  undelivered

and mounted on each nodes under /mnt/PBS/
and when i start pbs_mom on each node, i just specify the directory by
pbs_mom -d
pbs_mom -d /mnt/PBS/$HOSTNAME/PBS_spool

it's running fine.
[root at node001 ~]# ps aux |grep pbs
root     10760  0.0  0.0  10948  1036 ?        Ss   10:43   0:00
pbs_mom -d /mnt/PBS/node001/PBS_spool/


Yanan

On Wed, Jul 30, 2008 at 2:53 PM, Kamil Kisiel <kamil at zymeworks.com> wrote:
> Any reason why you recommend against writing the logs to NFS? We're looking
> at moving our logs off our nodes as well, since we're running a diskless and
> stateless configuration. Our /var/spool/torque directory is mounted in a
> tmpfs. Currently if a node reboots, we lose all of our mom_log, which is
> something I would like to avoid. I was thinking of doing a setup similar to
> what Yanan is doing, or perhaps just writing a program that would tail the
> mom log and dump the information to syslog (which, IMO, the MOM should be
> doing anyway).
>
> ____________
> Kamil Kisiel
> HPC Systems Engineer, Zymeworks Inc.
> 201-1401 West Broadway,
> Vancouver, BC, V6H 1H6, Canada
> Tel: (604) 678-1388 ext. 135
> Fax: (604) 737-7077
> www.zymeworks.com
>
>
>
> On 30/07/08 10:38 , "Glen Beane" <glen.beane at gmail.com> wrote:
>
> I have a local /var/spool/torque directory on each node and remotely mount a
> directory with the executables. I assume this is how almost everyone else
> does it.
>
>
> If you really want your logs all in one place then on each computer you
> could symlink /var/spool/torque/mom_logs to /raida/PBS/hostname/mom_logs
> where hostname is the name of the host you are creating the symlink on.  so
> you are setting up node10 you would ssh to node10 and symlink
> /var/spool/torque/mom_logs to /raida/PBS/node10/mom_logs
>
> you can also specify the path to a log file for pbs_mom on the command line,
> so if you want your logs written to your NFS mounted directory you could
> specify a path for the log file.  Note that if you do this you will not get
> a new log file every day, you'll have one log file that keeps getting bigger
> and bigger so you will want to setup log rolling (pbs_mom can do the log
> rolling if you set some parameters in the config file)
>
>
>
> I still think this is all a bad idea. I would have local directories for the
> mom logs though, I would not write the logs to NFS
>
>
>
> On Wed, Jul 30, 2008 at 1:24 PM, Yanan Sun <nancyprc at gmail.com> wrote:
>
> so you mean if i don't copy the whole thing on each computer, it can
> not have seperate jobs directory and log files?
>
> On Wed, Jul 30, 2008 at 1:15 PM, Glen Beane <glen.beane at gmail.com> wrote:
>> I would have a local pbs spool directory (like /var/spool/torque) on each
>> compute node.  I think it is a bad idea to mount them remotely.  You'll
>> run
>> into problems with sharing the mom_priv/jobs directory too I think.
>>
>>
>>
>>
>>
>> On Wed, Jul 30, 2008 at 1:03 PM, Yanan Sun <nancyprc at gmail.com> wrote:
>>>
>>> hi, there
>>>
>>> i have torque/maui installed in our cluster system. the headnode is
>>> running fine and i am trying to add more nodes. the files already
>>> mounted on each nodes, but for the way how it mounted, the mom_log
>>> file is read only. i don't want to change the way mount files on every
>>> nodes, soi am trying to change the directory of the log file goes to.
>>> now everything goes to /usr/PBS_spool/mom_logs, i want to change to
>>> /raida/PBS/$hostname/mom_logs
>>> how can i change it?
>>> thanks.
>>>
>>> Yanan
>>> _______________________________________________
>>> torqueusers mailing list
>>> torqueusers at supercluster.org
>>> http://www.supercluster.org/mailman/listinfo/torqueusers
>>
>>
>
>
> ________________________________
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
>
>
> ________________________________
> Notice of Confidentiality: The information transmitted is intended only for the
> person or entity to which it is addressed and may contain confidential and/or
> privileged material. Any review, re-transmission, dissemination or other use of
> or taking of any action in reliance upon this information by persons or entities
> other than the intended recipient is prohibited. If you received this in error
> please contact the sender immediately by return electronic transmission and then
> immediately delete this transmission including all attachments without copying,
> distributing or disclosing the same.
>


More information about the torqueusers mailing list