[torqueusers] Re: unable to contact node, Connection refused

[Alejandro Hurtado Turiño]

Hi Garrick and others, my situation go on, an now
grid1 is listed in the server_priv/nodes file.
I delete the "$restricted    grid1" line 
and MOM, SHED, and server are running.
[root at grid1 root]# ps ax|grep pbs
 2309 ?        S      0:19 /usr/local/sbin/pbs_mom
 2313 ?        S      0:00 /usr/local/sbin/pbs_sched
 2317 ?        S      0:00 /usr/local/sbin/pbs_server
 2889 pts/0    S      0:00 grep pbs
In fact, when I submit a job, this job is in queue, but dont start unless 
(like root) I send a command qrun job# then the jobs run successfully.
I don't configure a the pbs schedule it is by default, and I only start 
it.

thanks again,
Alejandro 


> 
> Thank Garrick, for your answer. but I continue with the problem and I 
> don't have any ideas on how to solve it. One time I made guide for the 
> installation of the torque/PBS server, based in my own(shortest) 
> experience and manuals. Now, I'm trying to to do it following this 
guide.
> (I write it at the end) 
> My server is a node too, and scheq (is it a problem? )

It's perfectly fine that the server is also a node.  Just make sure that
both pbs_server and pbs_mom daemons are running.

It's unclear to me whether the scheduler is running.  According to the
logs earlier, it wasn't.


> --------Configuracion server-sched 
> >torque.setup root                
> >echo 'grid1'>/usr/spool/PBS/server_priv/nodes   #creando el file nodes
> >gmgr -c 'create node worker2'        # and worker3,....

Nope.  'create node' is unsupported in qmgr.  You simply add the node to
server_priv/nodes and restart pbs_server.


> > > ---
> > > grid1 is the pbs server with pbsmon installed.
> > > no firewall
> > > my mom-priv/config
> > > $clienthost    grid1
> > > $logevent      255
> > > $restricted    grid1

Don't use $restricted unless you are on OSX with a broken bind()
function.

-- 
Garrick Staples, Linux/HPCC Administrator
University of Southern California