Hello Eygene,
Thanks for your feedback, I've changed to tcp till last friday and it seems to be working well, but we have also less alice's jobs...
Thanks again
Cheers
Christine
-----Message d'origine-----
De : Eygene Ryabinkin [mailto:[log in to unmask]]
Envoyé : vendredi 16 mars 2007 16:06
À : LEROY Christine
Cc : LHC Computer Grid - Rollout; HONORE Pierre-Francois SMTP; HONORE Pierre-Francois; DAPNIA/GRID_SUPPORT
Objet : Re: NFS hanging on WNs 4 jobs alice
Christine, good day.
> Till last Friday, we have a lot of alice jobs in our CE which is a good
> thing(CE: node07.datagrid.cea.fr ; VOBOX: node21.datagrid.cea.fr), but
> as we have 4 cores on our WNs, we have frequently 4 alice jobs in each
> WNs, which seems to cause problems with our mounted soft area by NFS:
> Some process are hanging indefinitely and I need to reboot the WNs (see
> end of email for such process)
> Has anybody already encountered this problem, and better, found a
> solution?
Once the JINR site had this problems. They moved from the UDP
to TCP NFS mounts and it solved the case.
--
Eygene Ryabinkin, RRC KI
|