[torqueusers] pbs_server init script?
nkadel at gmail.com
Mon Apr 5 05:57:06 MDT 2010
On Mon, Apr 5, 2010 at 7:41 AM, Peter Smith
<peter.smith3882100 at gmail.com> wrote:
> I am running Torque on Debian Lenny and i have been manually starting
> pbs_server and pbs_sched this is working fine. Now i have installed
> debian.pbs_sched and debian.pbs_server from the contrib/init.d folder
> and changed PBS_HOME to the correct folder. When running the script
> from a shell, when logged in as root, pbs_server starts fine but when
> the script is running as a part of the boot process the following
> error message is show:
Which version of Torque?
The pbs_server init script in 2.4.7 (which I looked at recently) has
an excess "fi" in it, where some shell scripting has been commented
out to avoid it accidentally re-initializing your scheduler.
> PBS_Server;Svr;PBS_Server;LOG_ERROR::process_host_name_part, host
> worker01 not found
> PBS_Server;Req;setup_nodes;could not create node "worker01", error = 15062
> PBS_Server;Svr;PBS_Server;LOG_ERROR::pbsd_init(setup_nodes), could not
> create node "worker01", error = 15062
> PBS_Server;Svr;PBS_Server;LOG_ERROR::PBS_Server, pbsd_init failed
> This error is occurring nine out of ten times when the server boots.
> When logging in and running the debian.pbs_server script manually
> afterwards pbs_server starts as expected. The debian.pbs_sched init
> script seems to be working perfectly as pbs_sched is always starting
> at boot without errors.
Hmm. Recent Linuxes have some very odd network management tools,
designed to "manage" their wireless or other network connections, but
which interfer with server class environments. I wonder if your
hostname, DNS and networking is fully up and running at the point at
which the pbs_server init script starts?
> Any suggestions on what could be wrong?
> torqueusers mailing list
> torqueusers at supercluster.org
More information about the torqueusers