Hi all,
Recently I've been trying to configure and run a cluster node on our
site (T2_HU_Budapest).
(Since we have 3 CE's it looks like it would be a good idea.)
I followed the available descriptions on EMI CREAM installation twiki,
and the YAIM twiki to setup the proper variable in CE's site-info.
I installed the emi-cluster package on one of the CREAM nodes,
then re-configured the CE in cluster mode (CE CREAM nodes are EMI-3
with SLC 6 OS).
I have now wrong publication according to lcg-infosites,
(gstat is unavailable so I couldn't check it there),
zero for all parameters, and 444444 as number of waiting jobs.
I found this issue in troubleshooting here
https://wiki.italiangrid.it/twiki/bin/view/CREAM/TroubleshootingGuide#4_2_Dynamic_information_not_publ
I ran manually the infoprovider script, but I found nothing suspicious
in its output. Although in
/var/log/bdii/bdii-update.log I have now errors as follows:
ERROR:lcg-info-dynamic-scheduler:Execution error: Error running
"/usr/libexec/glite-info-dynamic-sge --vomaxjobs -c
/etc/lrms/scheduler.conf": OSError(2, 'No such file or directory')
Since I haven't found any thorough and consistent install/config guide
concerning the emi cluster,
I suspect that there might be some misunderstanding on my part
regarding the cluster mode configuration.
Therefore if someone has experience with cluster mode, and also could
share it with me,
it would be greatly appreciated.
Cheers,
Andras
|