...
- Before admin-suspending jobs, it is recommended to disable scheduling and wait for the current scheduling cycle to finish. The scheduler only queries the vnode state at the start of the cycle. If a vnode moves into 'maintenance' after the cycle starts, the scheduler may still consider the vnode as schedulable. It is possible for new jobs to start during the current cycle.
- If an admin wants to perform maintenance on a vnode that has no jobs running on it, they should put the vnode in the offline state and perform maintenance.
- Any reservations on vnodes in the maintenance state will be marked degraded. PBS will search for alternate vnodes for the reservations.
- Sub jobs are requeued upon server restart. Any vnode which only had admin-suspended subjobs will return to the free state after a server restart.
- As with all pseudo-signals, the new ones do not have a signal number associated with them. Signal numbers are OS defined signals. Pseudo-signals are PBS constructs that are special cases.
- If a job is running on some but not all of the vnodes of a multi-vnoded host, only the vnodes the job is running on will be put into maintenance.
- It is suggested that all jobs on all vnodes of a muilti-vnoded host be admin-suspended before starting maintenance. If not, some vnodes may remain in a schedulable state and have new work started on them during maintenance.
...