How-to view Talapas compute usage
- Lucas Crownover
This page describes how to view Talapas service units consumed. Talapas usage is now displayed on login, following the MOTD, e.g.:
Welcome to Talapas!
...Message-Of-The-Day...
Talapas Usage for user ryelle in hpcrcf PIRG as of midnight 2018-08-09:
hpcrcf Compute (SU): ryelle: 16 Group: 166 Remaining: 999834
hpcrcf Storage (GB): ryelle: 1482 Group: 10083 Quota: 15360
[ryelle@talapas-ln2 ~]$
This says that user "ryelle" has consumed 16 SUs so far in FY2019 (start July 1 through the end of the day on August 9 in this case) and his PIRG has consumed 166 SUs, and there are 999834 SUs remaining in his PIRG's (projected) platinum compute membership. The message also displays the storage currently used by user ryelle (1482 GB) and the hpcrcf PIRG (10083 GB), as well as the storage quota for the hpcrcf PIRG (15360 GB) from their projected gold storage membership. Home directory storage is currently not displayed. The service units and storage are computed and updated at the end (12am) of every day.Â
The using the "report" command can be used to obtain usage information on specific jobs, or over specific time frames, or more up-to-date usage. Here are some examples:
1) Get help on using the report command:
[ryelle@talapas-ln2 ~]$ report -h
Report service units consumed:
-j|--jobs= : which jobids (default is all in given time frame)
-S|--start-time= : Start time (e.g. YYYY-MM-DD, default is current day)
-E|--end-time= : End time (e.g. YYYY-MM-DD, default is now)
2) Default behavior is to obtain usage for all jobs for the current day until the present (in this case no jobs have been run):
[ryelle@talapas-ln2 ~]$ report
No Service Units consumed over the time frame requested
3) Display service units consumed by all jobs run in the month of July, 2018:
[ryelle@talapas-ln2 ~]$ report -S 2018-07-01 -E 2018-08-01
# for date range 2018-07-02T09:17:24
# to 2018-07-31T15:41:21
#
# account/user Grand-SU CPU-SU Mem-SU GPU-SU
hpcrcf 15.7 14.8 0.9 0.0
hpcrcf - ryelle 15.7 14.8 0.9 0.0
4) Display service units consumed by a specific job (e.g. 2090008):
[ryelle@talapas-ln2 ~]$ report -j 2090008
# for date range 2018-07-30T19:34:22
# to 2018-07-30T19:34:22
#
# account/user Grand-SU CPU-SU Mem-SU GPU-SU
hpcrcf 8.0 8.0 0.0 0.0
hpcrcf - ryelle 8.0 8.0 0.0 0.0
For an explanation of how Service Units are computed, visit this link: