Hi Rod,
we have same problem before but i cant remember the links generated with
simple fork testing, but i am sure the logmsg did contains lots of sanity
testing statement. this is becuase we're trying to adjust the gridmap and
groupfile as well as voms-grid-mapfile to be able to fit into current
backend batch system. after that, thought we dont have problem with job
submission (via job manager lcgpbs or fork) but the GK log tend to
increase rapidly. this cause system root full in less than one month. i
try updating the users.conf and group config under yaim and reconfigure
again, since then, we dont have same problem and content of GK log back to
normal entries wrt all job submission. this have been discuss in rollout
earlier but i cant remember the details including also the input from
Maarten earlier. fyi.
Br,
J
On Thu, 22 May 2008, Rod Walker wrote:
> I restarted the gatekeeper and did one globus-job-run
>
> [root@ce1 ~]# wc -l /var/log/globus-gatekeeper.log
> 19134984 /var/log/globus-gatekeeper.log
> [root@ce1 ~]# wc -l /var/log/globus-gatekeeper.log
> 19142376 /var/log/globus-gatekeeper.log
>
> which generates 8000 lines.
>
> Does anyone else see this? Maybe it`s tolerated because it zips to almost
> nothing.
>
> Cheers,
> Rod.
>
> On Fri, 23 May 2008, Maarten Litmaath wrote:
>
> > On Thu, 22 May 2008, Rod Walker wrote:
> >
> > > I have
> > > gridenv_set "LCMAPS_LOG_LEVEL" "1"
> > > gridenv_set "LCMAPS_DEBUG_LEVEL" "0"
> > > gridenv_set "LCAS_LOG_LEVEL" "1"
> > > gridenv_set "LCAS_DEBUG_LEVEL" "0"
> > >
> > > These seem to be the defaults so the only improvement would be to set
> > > *_LOG_LEVEL=0 - is that wise?
> >
> > No, the aforementioned values are OK.
> > Did you try restarting the gatekeeper?
> >
> >
>
>
|