OS Version: CentOS release 4.6 (Final) Kernel \r on an \m 2.6.9-100.ELsmp
When I attempt to run a job it gives me the error as follows.
qsub: Bad UID for job execution
I have created a fresh user account and the same error occurs, yet other users on the same machine can run jobs without a problem.
How can I fix this issue?
Instructions here: http://www.bo.infn.it/alice/introgrd/pbsabout/node18.html suggest that you are running it as wrong user.
Invalid user in #PBS -u new-user $ qsub sube.cmd qsub: Bad UID for job executionThe user name defined in the user list is either undefined or different from the user name of the job submitter or the user UID and GID at the executing node are different from the ones in the submitting node. In this case, first of all check that the #PBS -u directive specifies a valid user, then check that user IDs are the same:
$ id
whoamiuid=501(myusrname) gid=501(myusrname) groups=501(myusrname) $ grep -e "PBS -u" cpbs.cmdPBS -u alice001 $ id alice001 uid=10417(alice001) gid=1395(alice) groups=1395(alice)
and change the #PBS -u as appropriate.
Adding the host the /etc/hosts.equiv fixed my issue.
But I had already added these lines to qmgr.
set server operators += [email protected]
set server acl_users += username
set server managers += [email protected]
set server acl_hosts += servername.local
set server operators += [email protected]
Then I did a /etc/init.d/pbs_server restart.
You can restart the pbs_server even if you are already running.
You don't have to add the server to /etc/hosts.equiv if you are submitting jobs on the server that is running the pbs_server.
I know this question is a bit old, but I recently ran into this issue and I performed the following steps:
qmgr -c "set server flatuid=true"qmgr -c "set server acl_roots+=username@*"qmgr -c "set server operators+=username@*"Here * is a wildcard for all hosts. Works like a charm! Probably not super secure but definitely works!
In my scenario, I received this error when trying to submit a job while working in an interactive job. Exiting the interactive qsub and submitting from the head node worked.
two procedures:
add ""+::::::"" at the end of the /etc/passwd file
make sure the node where users submit jobs have all other nodes well defined in file /etc/hosts.equiv
reference: search Bad UID for job execution in torque admin guide pdf.
