Error messages and other odd happenings (or non-happenings)

The systems you use are often complex, and many individual computers and processes on them have to all be running correctly for the expected outcome. Sometimes if a vital link in this process is missing, you get error messages. By tradition these are meant to be cryptic, terse, and uninformative. But there are often explanations. This page strives to demystify some of these error messages and to explain what might be happening (or not happening) under the hood.

Unable to start job

This is produced by the HPC portal when you try to submit a job to the queues. The portal tries to contact the batch scheduling software (PBS) running on a master node for the HPC system in order to schedule the job. If that node, or the PBS process on it is not contactable, you get this error message. It takes about 10 minutes to reboot the vital systems, so you can try again in say 20 minutes to see if the HPC has been rebooted. If the delay is longer, it might mean a significant hardware failure somewhere, or even power outages.