I have installed pbspro on a brand new cluster where Centos 7.4 is installed.
I built the rpms and installed the server one on the frontend and the execution one on a node.
Now I am trying to configure a first test queue, but any command I execute in qmgr ends with;
qmgr obj=<—something—> svr=default: Unauthorized Request
qmgr: Error (15007) returned from server
I am launching qmgr as root; but the root user should be a manager by default, correct ? Am I missing something ?
Thank you for any assistance,
I am running the command on the server itself
I have two entries in /etc/host for the server, the “internal” and the “external” one;
192.168.33.1 werewolf.werewolf werewolf
I have tried setting PBS_SERVER to each of these values but the error still occurs
Yes, werewolf is resolvable.
For some reason , the PBS Server thinks, the commands originating from itself is coming from a different host. Hence, you are seeing this message qmgr: Error (15007) returned from server
If you open your server logs, you can see what the server thinks its hostname is:
08/27/2019 08:57:34.688203;0040;Server@shecil;Svr;shecil;Scheduler sent command 3
In this case, the server thinks it’s running on shecil (Server@shecil)
If you run a command that connects to the server, the server longs will print the hostname of the incoming request:
[vstumpf@shecil ~]$ qstat
[vstumpf@shecil ~]$ tail -n 4 /var/spool/pbs/server_logs/20190827
08/27/2019 09:10:09.693158;0100;Server@shecil;Req;;Type 0 request received from vstumpf@shecil, sock=15
In this case, the request is coming from vstumpf@shecil. The server will do a string comparison of the two hosts.
If I issue a “qstat” command or I try any qmgr command nothing is written in the server log. Do I need to enable the logging somewhere ?
Moreover, digging into documentation I found this troubleshooting suggestion
Additionally, try running pbs_iff by typing
pbs_iff server_host 15001
If pbs_iff fails to print the garbage and/or exits with a nonzero status, either the Server is not running or it was installed with a different encryption system from that used for pbs_iff.
The command exists with zero status, but nothing is printed; could this be the problem ?
Logging is enabled by default, so no configuration needs to be done.
qmgr -c "set server log_events = 511 "
0. please check all your PBS Pro OSS commands are the same version as the PBS Server
0. please check chmod 4755 $PBS_EXEC/sbin/pbs_iff
- Do you have any jobs running in your PBS setup ?
- try deleting a node and adding a node
- create a queue and delete a queue
[root@admin ~]# pbs_iff -t admin 15001 ( i do not get anything printed out after running this command )
[root@admin ~]# echo $?
It seems that the
chmod 4755 $PBS_EXEC/sbin/pbs_iff
did the job!
Even if the setuid permission on this file seemed to me already setup before… anyway, now it works.
Thank you very much for the assistance,