Information for Group Managers

A group manager is a person authorized to manage group course or private cluster course.
By default, the group manager is set to the service course applicant. The applicant of the service course can add/change the group manager. Please apply for add/change via User Portal.

A group manager can use the commands for managing the group. The dedicated commands of group manager allow group managers to manage queues and disks allocated to their groups and group members.

Group members can be managed by logging into the User Portal.

  1. After logging in to the User Portal, click “Service Application” on the top menu group_en_1
  2. Click “Group members” on the left menu. group_en_2
  3. You can check and add/delete members from “Member management”. group_en_3

You can make backup settings of the LARGE disk space by using the group_backup command. LARGE disk space consists of the /LARGE0/groupname directory and the /LARGE1/groupname directory, and you can set one of the following status.

Status /LARGE0/groupname /LARGE1/groupname
Backup Safe（Make backup） Backup （Backup location）
Not Backup Unsafe（Not make backup） Unsafe（Not make backup）

Of these, disks whose settings are Safe or Unsafe can be used.

Checking the backup settings

The target group can be specified with the -g option. If omitted, the current group when the command is executed is targeted.

$ group_backup -g gr19999 -l 
 Num  Filesystem             Status  Filesystem             Status
  1)  /LARGE0/gr19999  ... Safe    /LARGE1/gr19999    ... Backup　<- バックアップ使用状態

Setting the status to “Not Backup”

$ group_backup -g gr19999 --unsafe 1
/LARGE0/gr19999: Safe   => UnSafe
/LARGE1/gr19999: Backup => UnSafe

Checking the backup settings（after changes）

$ group_backup -g gr19999 -l
 Num  Filesystem             Status  Filesystem             Status
  1)  /LARGE0/gr19999    ... UnSafe  /LARGE1/gr19999    ... UnSafe  <- バックアップ未使用状態

Return to the status to Backup

$ group_backup -g gr19999 --safe 1
/LARGE0/gr19999: Unsafe => Safe
/LARGE1/gr19999: Unsafe => Backup

The group_trash command allows users to delete files(move to trash) in the LARGE disk space. It can delete the data files of users who are no longer enrolled due to graduation, etc. If you accidentally delete a file, you can recover it from the trash, but please note that the trash is emptied every Monday.

Deleting files by the group_trash command

Specify the target group with the -g option. If omitted, the group manager authority are determined by the current group when the command is executed.

$ group_trash -g gr19999 /LARGE0/gr19999/file1 
file1 to Trash (/LARGE0/gr19999/.DpcTrash/b59999/2009-04-10_1010)

There are two types of units of management of Slurm queue privileges: users and groups. Users are empty by default, Groups are initially registered with the group corresponding to the queue name as the default setting.

If you wish to use a queue with multiple groups, or if you wish to grant queue access to a single user who does not belong to a group, please contact us using the Inquiry Form.

You can select the job scheduling policy from the following three types. If you are using an individual queue (grXXXXXx), you can change your preferred scheduling policy by contacting us from the applicant at Inquiry Form. (We cannot accept requests for shared queues such as entry course or personal course.)

Settings Operation
pass If there are sufficient computing resources to execute a job, it will overtake jobs waiting to be executed that are in line before it.
You can use computing resources efficiently, but large jobs may not be executed indefinitely.【Default Value】
wait It will not overtake jobs even if there is a enough computing resources.
backfill Based on the calculations of the execution time limit (-t) of each job, it will overtake jobs only if it does not affect the execution start time of other jobs. For example, you can use resources effectively by executing small jobs that can complete execution before a large job is started.

You can confirm the job execution status of queues registered as a group manager and cancel the jobs with the spadmin command.

  • Confirmation of job execution status 

    $ spadmin list -p gr19999b ## Please change the "gr19999b" part to the queue name you wish to confirm.
         JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
          4781  gr19999b run_cpu2   b59999  R    1:26:09      1 nb0001

  • Cancellation of the job

    $ spadmin cancel 123
scancel: Terminating job 123

If anyone other than the group manager execute the spadmin command, the following error message will be displayed.

$ spadmin list -p gr19999g
Authorization Failure