Issue with qrsh (GE)

Ansys Electromagnetics parallel batch jobs use the GE qrsh command to launch engine processes on remote hosts. If the qrsh command is not working correctly, then the parallel job is unable to launch engine processes on remote hosts. If this problem occurs, the batch log for the job typically includes one or more error messages indicating that a COM engine was unable to be started on a remote host. If this occurs, the user or cluster administrator should verify that the GE qrsh command is working correctly, and correct the problem if the GE qrsh command is not working correctly.

The qrsh command may be tested by running a simple command on a specified host, such as qrsh -l hostname=host1 hostname or qrsh -l hostname=host1 ls /tmp, where host1 is the remote host name. The first test should return the hostname of the remote machine. The second test should list the contents of the /tmp directory on the remote machine.

The failures of the GE qrsh command are associated with the following global GE configuration parameters, listed below with values that may cause the failures:

qrsh_command /usr/bin/ssh -t

rsh_command /usr/bin/ssh -t

rlogin_command /usr/bin/ssh -t

If these parameter settings are removed, then the GE built-in mechanisms are used for qrsh, rsh, and rlogin. No problems with the built-in versions have been reported. The GE qconf -sconf global command may be used to view these parameter settings. The GE qconf -mconf global command may be used to modify or remove these parameter settings.