Error: pbs_iff: cannot read reply from pbs_server

Question

I have installed Torque-4.2.6 version. My machine has RHEL 6 Operating System. I have installed through user1 login as it has root permission. It is working properly whenever I execute 'pbsnodes -a' command. If I'm login through user2 onto the same machine, whenever I execute 'pbsnodes -a' then it is throwing the following error.

pbs_iff: cannot read reply from pbs_server No Permission. pbsnodes: cannot connect to server 'node1', error=15007 (Unauthorized Request )

There no such executable as pbs_iff. What could be the problem.? Can anyone help me.?

Answer 1

Here is my experience with PbsPro, I would assume it works similar with Torque.

What you need is to change the pbs_iff mod to -rwsr-xr-x by:

chmod 4755 /opt/pbs/.../sbin/pbs_iff

Refer to the location of pbs_iff file in your system. Note it is in the pbs.../sbin folder instead of the bin folder.

Answer 2

Per https://serverfault.com/questions/253932/torque-works-half-of-the-time-fails-no-permission-the-other-half -- there may be an issue with the queue being clogged with corrupt jobs, possibly from a dead node, so manually clearing out

/var/spool/torque/server_priv/jobs/

and

/var/spool/torque/server_priv/arrays/

and restarting pbs_server may solve the issue.

Note you will obviously have to resubmit these jobs back to the queue

Answer 3

qmgr

Qmgr:set set server acl_host_enable=False