|
|
|
BGL scheduling policy
The current BGL job scheduling policy is very simple and is currently managed
in part by the resource manager.
- 32 node partitions are active at all times except when a reservation is in
place.
- We currently have one (1) 64 node partition that is active at all times except when a reservation is in place.
- 1024 node jobs may be run at any time during the night or weekend, as long
as a reservation for the full rack has been put into place. We ask that you
send us reservation requests early in the week (or the week before) so we can plan
the week's reservations.
- 512 node jobs may be run at any time during the day or night, as long as a
reservation for the partition has been put into place. Shorter (< 30 minutes)
512 node jobs may run without a reservation if the resources have not been
reserved by anyoone else.
- To get a 512 node reservation, please send email to
support@bgl.mcs.anl.gov with the desired start time and duration. If
possible, a reservation for that time period will be put in place (if
there are any 32 node jobs running in the 512 node partition, we will
contact the owner and ask them if we can kill them, etc). If not, you
will be notified and asked to pick a later time. Putting a reservation in
place means that we tell the scheduler that at time T, 32 node partitions
X,Y,Z,... are reserved and 512 node partition 'A' becomes active.
As the time for a reservation approaches. the scheduler will not run 32
node jobs in the set of reserved partitions X,Y,Z... that would run into
the reserved time period.
Once the reservation start time arrives, any 512 node jobs will be
scheduled in partition 'A', FIFO. We are hoping to have 'owned'
reservations shortly. Then, only jobs owned by the person(s) with the
reservation will run.
General Scheduling Guidelines
BGL is expected to support all kinds of different usage, we do not
want to unnecessarily limit the kinds of work that could be done on
BGL. We would prefer to maintain a flexible scheduling policy based
on the needs of the community rather than impose a strict policy.
We ask that all users follow these guidelines:
- Follow good etiquette. This means, among other
things, don't start jobs that use more than 64 nodes
except during a reserved time.
- Clearly, some work will require use of the machine beyond those
bounds. In those cases, please send a quick note to
support@bgl.mcs.anl.gov so that we can make arrangments with the
users to accomodate you.
- If we get well-founded complaints from other users of the
system about your jobs, we will attempt to contact you to determine
the best course of action. However, under some circumstances, we
may have to kill running jobs.
|
|