[GE users] Re: submit jobs to specific hosts

Nano Surbakti nano.surbakti at gmail.com
Thu Aug 31 06:25:00 BST 2006


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

sge_execd is still running
I check the message in sge-root/default/spool/qmaster:

08/31/2006 12:09:39|qmaster|elka-70|I|read job database with 0 entries
in 0 seconds
08/31/2006 12:09:39|qmaster|elka-70|I|qmaster hard descriptor limit is
set to 8192
08/31/2006 12:09:39|qmaster|elka-70|I|qmaster soft descriptor limit is
set to 8192
08/31/2006 12:09:39|qmaster|elka-70|I|qmaster will use max. 8172 file
descriptors for communication
08/31/2006 12:09:39|qmaster|elka-70|I|qmaster will accept max. 99
dynamic event clients
08/31/2006 12:09:39|qmaster|elka-70|I|starting up 6.0u8

Another info:

I can't do rsh from master host (named elka-70) to the exec host, so I
can't do the job submit test in "N1 Grid Engine 6 Installation Guide"
(817_6118.pdf, page 86).

When you said "the sge_execd spool directory for the node", I think I
don't have it. In the exec node the $SGE_ROOT is not owned by
sgeadmin. The permission look like this:

drwxr-xr-x  18 507      root        4096 Aug 18 00:26 sge-root

Perhaps I need to do something with NFS setting? The sgeadmin is exist
both on exec node and master node, with the same password. Do I need
to setup NIS to sync the account?

There are so much new things for me here. Sorry, if I didn't
understand many things.

Thanks for your help.

--
Nano Surbakti

On 8/30/06, Sean Dilda <sean at duke.edu> wrote:
> Nano Surbakti wrote:
> > On 8/29/06, Davide Cittaro <davide.cittaro at ifom-ieo-campus.it> wrote:
> >>
> >> Sorry for the stupid question, but is sge_execd running on nodes?
> >
> > Yup.. it's running. I have tested it to run a simple job.
> >
>
> It sounds like sge_execd was communicating with sge_qmaster, thens
> topped.  Please check that sge_execd is still running.  Please look in
> the 'messages' files in the sge_qmaster spool directory, and the
> sge_execd spool directory for the node.  Those files often have useful
> information when something goes wrong.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list