[GE users] NFS errors

Thomas Dehn - PTS Software Engineer Thomas.Dehn at Sun.COM
Thu Apr 21 14:40:13 BST 2005


It is the shepherd which tries to write to the
trace file. In theory, the shepherd can write
as long as root has write access.

Thats where the nfs configuration comes into
play. Per default, the nfs client's root only has "nobody" access
rights to a file mounted via nfs. And there exist oddities
in certain vendors' nfs implementations which, and sometimes
even intermittendly, prevent the nfs client's root
from accessing the files even when theoretically
root should have root access rights.

If a local spool directory is acceptable to you,
thats probably a much faster solution than pinning
down why the nfs access fails.


With kind regards,


Thomas Dehn


-- 
*******************************************************
Thomas Dehn                     Tel: (+49 89) 46008 2856
PTS Software Engineer		Fax: (+49 89) 46008 2777
Sun Microsystems GmbH           Mobile: 
Sonnenallee 1                   mailto: thomas.dehn at sun.com
D-85551 Kirchheim-Heimstetten   http://www.sun.de 





From: Schmitz Dale M Contr 20 IS/INPG <Dale.Schmitz at offutt.af.mil> wrote:
> 
> For what it's worth, and I may be jumping into the middle of a
> conversation here, I had the very same errors two weeks ago.  They were
> due to my SGE installation directory (SGE_ROOT) being NFS served.  A
> subdirectory of that, $SGE_ROOT/default/spool/<hostname>, needed to be
> local.  So I made a link to /var/tmp/<hostname>.  The active_jobs
> directory is now on a local file system and all my NFS errors have gone
> away.
> 
> Dale
> 
> -----Original Message-----
> From: Thomas Dehn - PTS Software Engineer [mailto:Thomas.Dehn at Sun.COM] 
> Sent: Thursday, April 21, 2005 4:16 AM
> To: users at gridengine.sunsource.net
> Subject: Re: [GE users] NFS errors
> 
> *Timo Viitanen <Timo.Viitanen at csc.fi> wrote:
> > >     Do you mean in the /var/adm/messages and something like this:
> > > 
> > >     Apr 12 01:40:00 corona1 nfs: [ID 897781 kern.notice] NFS write
> error
> > >     on host corona-sge-g: Permission denied.
> > >     Apr 12 01:40:00 corona1 nfs: [ID 702911 kern.notice] (file
> handle:
> > >     3799a28 3e7 a0000 5a1 71f57435 a0000 2 7744)
> > >     Apr 12 01:40:00 corona1 nfs: [ID 897781 kern.notice] NFS write
> error
> > >     on host corona-sge-g: Permission denied.
> > >     Apr 12 01:40:00 corona1 nfs: [ID 702911 kern.notice] (file
> handle:
> > >     3799a28 3e7 a0000 5a1 71f57435 a0000 2 7744)
> > >     Apr 12 01:40:00 corona1 nfs: [ID 897781 kern.notice] NFS write
> error
> > >     on host corona-sge-g: Permission denied.
> > >     Apr 12 01:40:00 corona1 nfs: [ID 702911 kern.notice] (file
> handle:
> > >     3799a28 3e7 a0000 5a1 71f57435 a0000 2 7744)
> > > 
> > >     Or in the default/spool/<hostname>/messages and something like
> this:
> > > 
> > >     04/14/2005 12:36:14|execd|corona1|W|reaping job "6269" ptf
> > >     complains: Job does not exist
> > >     04/14/2005 12:36:14|execd|corona1|E|can't open usage file
> > >     "active_jobs/6269.1/usage" for job 6269.1: No such file or
> directory
> > >     04/14/2005 12:36:14|execd|corona1|E|"can't read usage file for
> job
> > >     6269.1
> > >     "
> > >     04/14/2005 12:38:00|execd|corona1|W|reaping job "6270" ptf
> > >     complains: Job does not exist
> > >     04/14/2005 12:38:00|execd|corona1|E|can't open usage file
> > >     "active_jobs/6270.1/usage" for job 6270.1: No such file or
> directory
> > >     04/14/2005 12:38:00|execd|corona1|E|"can't read usage file for
> job
> > >     6270.1
> > > 
> > >     I'm running N1GE6U3 and Sun support doesn't have a clue where
> they
> > >     are coming from... (Service ticket has been open for last
> 4-5months)
> 
> I am very sorry this takes so long, but until recently
> I did not receive important information I first requested 
> back on november 25th, and requested a few more times since then. 
> It was my impression that you were too busy with more important 
> issues to provide the requested data. We should have been more
> proactive with you in this case but we will do our best in the
> future to revisit dormant support cases.
> 
> If you wish to discuss this further, let us please take this discussion
> off the public alias and continue the conversation in private.
> 
> 
> With kind regards,
> 
> 
> Thomas Dehn
> 
> 
> -- 
> *******************************************************
> Thomas Dehn                     Tel: (+49 89) 46008 2856
> PTS Software Engineer		Fax: (+49 89) 46008 2777
> Sun Microsystems GmbH           Mobile: 
> Sonnenallee 1                   mailto: thomas.dehn at sun.com
> D-85551 Kirchheim-Heimstetten   http://www.sun.de 
> 
> 
> 
> 
> ---------------------------------------------------------------------
> 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
> 


---------------------------------------------------------------------
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