Change Details

In pmtpa, we limited the number of jobs that could be executed in parallel per queue to 16 IIRC. During migration to eqiad, this seems to have been lost, and thus at the moment tools.currentevents has 86 jobs running (further only queued because the load of the exec nodes is saturated): | scfc@tools-dev:~$ qstat -u tools.currentevents | fgrep ' r ' | wc -l | 86 | scfc@tools-dev:~$ So we need to limit the number of jobs executed in parallel again. Last time, there was some confusion what configuration option limited what and that initially caused only to limit the number of /pending/ jobs and delete others, so we need to be careful about that. -------------------------- **Version**: unspecified **Severity**: normal

In pmtpa, we limited the number of jobs that could be executed in parallel per queue to 16 IIRC. During migration to eqiad, this seems to have been lost, and thus at the moment tools.currentevents has 86 jobs running (further only queued because the load of the exec nodes is saturated): ``` scfc@tools-dev:~$ qstat -u tools.currentevents | fgrep ' r ' | wc -l 86 scfc@tools-dev:~$ ``` So we need to limit the number of jobs executed in parallel again. Last time, there was some confusion what configuration option limited what and that initially caused only to limit the number of /pending/ jobs and delete others, so we need to be careful about that.

In pmtpa, we limited the number of jobs that could be executed in parallel per queue to 16 IIRC. During migration to eqiad, this seems to have been lost, and thus at the moment tools.currentevents has 86 jobs running (further only queued because the load of the exec nodes is saturated): | scfc@tools-dev:~$ qstat -u tools.currentevents | fgrep ' r ' | wc -l``` | 86scfc@tools-dev:~$ qstat -u tools.currentevents | fgrep ' r ' | wc -l | scfc@tools-dev:~$ So we need to limit the number of jobs executed in parallel again. Last time, there was some confusion what configuration option limited what and that initially caused only to limit the number of /pending/ jobs and delete others, so we need to be careful about that. --------------------------86 **Version**: unspecifiedscfc@tools-dev:~$ **Severity**: normal``` So we need to limit the number of jobs executed in parallel again. Last time, there was some confusion what configuration option limited what and that initially caused only to limit the number of /pending/ jobs and delete others, so we need to be careful about that.