I've run into a configuration problem when I enable GSI authentication between my master and the worker node(s). The jobs I submit are executed, the logs are written out fine, but the job is never removed from the queue.
The ShadowLog contains messages like this:
01/24/14 00:40:23 (64.5) (22215): SECMAN: could not receive session info, failing!
01/24/14 00:40:23 (64.5) (22215): Can't connect to queue manager: SECMAN:2007:could not receive post_auth_info.
01/24/14 00:40:23 (64.5) (22215): Failed to perform final update to job queue!
for each job sitting in the queue. Any idea why the queue does not get updated after the cleanup?
University of Victoria
Physics & Astronomy
UVic Phone: +1 (250) 472-4085
UVic Office: Elliot 212