On 11/25/2015 07:49 PM, Maarten Litmaath wrote:
> Hi Adrian,
Hi!
>>> you are probably missing the lrms_backend_cmd in your
>>> /etc/lrms.scheduler.conf:
>>>
>>> [LRMS]
>>> lrms_backend_cmd: /usr/libexec/lrmsinfo-pbs -s torque.farm.particle.cz
>> thanks a lot! this was indeed the case .. but i wonder why this
>> happened? did something changed in new packages?
>
> Maybe you did not (fully) reconfigure the CE after some rpms got updated?
yeah, i did not .. i was kind of hoping that when rpms are updated the
configuration in place should not be reset (like the all other updates
that creates a conf.new file if there is a previous file in place .. )
i am curious about why there is reconfiguration needed after updates ..
i mean, if the new package brings new necessary defaults maybe it would
make sense but i would think that the existent configuration, no matter
the service, should not be touched by rpm installation procedure ..
(AFAIK this is the standard for all linux distros)
> Note that service nodes should not update their MW rpms automatically,
> because a reconfiguration and/or a preparation may be needed.
yeas, i am painfully aware of this (i almost got bitten by epel torque
problem) so i dont have automatic updates enabled
Thanks!
Adrian
|