Re: [Condor-users] condor_submit hangs / condor_q hangs

[Marco Mambelli <marco@xxxxxxxxxxxxxxxx>]

Hi Adam,
are the 2 submit nodes behaving the same way?
did you set CONDOR_HOST to point to your "master node"?
Are you running master, collector and negotiator on the "master node"?

Depending on the configuration you may have logs also in /tmp

scedd is single threaded, if submission is hanging then also condor_q is 
hanging.

Marco



On Thu, 18 Mar 2010, Steven Timm wrote:

> What's the content of SchedLog on the submit machine, sounds
> like it could be some kind of authentication issue between
> condor_submit, condor_q and the schedd, or else a schedd that's
> just totally hosed for some reason.
>
> also you can
>
> export _TOOL_DEBUG=D_ALL ; condor_submit -debug <args>
> and you can get some debugging info from condor_submit.
>
> Steve
>
>
> On Thu, 18 Mar 2010, Adam Yates wrote:
>
> >  Hi everyone;
> >
> >  I'm having a problem with a fresh condor install.
> >  My setup is this:
> >
> >  1 master node (master)
> >  2 interactive nodes (submit only- n00 and n01)
> >  64 worker nodes (execute only n02-n66)
> >
> >  Config structure:
> >
> >  master - /home/condor is nfs exported to all nodes. local configs are in
> >  /home/condor/$HOSTNAME/condor_config.local
> >
> >
> >  Whenever I use condor_submit, it hangs on "Submitting job.." and then
> >  eventually times
> >  out, saying that it failed to connect to the  local machine on port x. and
> >  failed to fetch
> >  ads from the localhost on port x.
> >
> >  Whenever I use condor_q, it will hang but if I give condor_q -global, it
> >  will return the
> >  status of my condor pool and show some of the nodes in use.
> >
> >  The daemon listing in my config on the submit nodes is as such:
> >  DAEMON_LIST = MASTER, SCHEDD
> >
> >  There are no errors in the local log (log/*)
> >
> >  Does anybody make sense of this? Please let me know if any more info is
> >  needed.