[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] Incorrect slot resources after job is vacated



On Mar 17, 2014, at 11:18 AM, Benjamin Tovar <btovar@xxxxxx> wrote:

We are trying to diagnose a problem with slots and requested resources after a
job is vacated.

For example, if we set:

request_cpus = 1

in the submit file, the resulting job does get allocated to a 1 core slot.
However, if the job is vacated, when it reappears it does so in a 4 cores slot.
When vacated, the now 4 cores job remains at 4 cores. This happens every time a
job is vacated.

We appreciate any hints on where to look to solve this issue.

Is it possible that there just happen to be a lot of 4-core claims being reused to run this 1-core job? If the same user has a lot of 4-core jobs, that could happen.

Thanks and regards,
Jaime Frey
UW-Madison HTCondor Project