Command Hints

The page contains a collection of command hints. These are commands that you might find useful for determining the status of jobs on RA.

showbf

The showbf command allows users to determine what resources are available for immediate usage. That is, how many nodes are available at the current time to run a job and for how long. If you submit a job for less than the shown amount of time and number of nodes it should start immediately.

Example:

		
[tkaiser@ra ~]$ showbf 
Partition     Tasks  Nodes   StartOffset      Duration       StartDate
---------     -----  -----  ------------  ------------  --------------
ALL              24      3      00:00:00      00:59:57  13:04:34_09/15
ALL              16      2      00:00:00      INFINITY  13:04:34_09/15

This shows that I could start a job on 3 nodes with a run time of under 1 hour and have it run immediately.

showstart

Get the esstimated start time of job, where ##### is the jobid

Example: get the esstimated start time of job 83862.

[tkaiser@ra ~]$ showstart  83862
job 83862 requires 24 procs for 1:06:00:00

Estimated Rsv based start in                 2:51:33 on Tue Sep 15 16:06:35
Estimated Rsv based completion in         1:08:51:33 on Wed Sep 16 22:06:35

Best Partition: base

[tkaiser@ra ~]$

checkjob

Checks the status of one of your jobs, where ##### is the jobid

Example: check the status of job 79042.

[tkaiser@ra ~]$ checkjob 79042
job 79042

AName: DFT_config6
State: Idle 
Creds:  user:tkaiser  group:tkaiser  account:challenge  class:MEDIUM2
WallTime:   00:00:00 of 2:00:00:00
SubmitTime: Mon Aug 24 16:45:18
  (Time Queued  Total: 17:51:09  Eligible: 16:25:34)

Total Requested Tasks: 32

Req[0]  TaskCount: 32  Partition: ALL  
Memory >= 0  Disk >= 0  Swap >= 0
Opsys:   ---  Arch: ---  Features: ---


BypassCount:    12
Partition List: base
StartPriority:  10219
rejected for State        - 
rejected for Reserved     - 
NOTE:  job cannot run in partition base (idle procs do not meet requirements : 0 of 32 procs found)
idle procs:  32  feasible procs:   0

Node Rejection Summary: [State: 264][Reserved: 4]


[tkaiser@ra ~]$

This shows that job 79042, with name DFT_config6 is currently not running, (Idle) It was submitted by user tkaiser who is in the challenge accounting group. See:http://petra.mines.edu/cgi-bin/rastats/index.cgi for a listing of machine usage by accounting groups.

2 days were requested for this job on 32 processors.

It is currently not running because 32 free processors were not found the last time the scheduler checked for available nodes. Note: Other 32 node jobs might run first if they have a higher priority or if they are submitted for a short time.

mdiag -p

Get the priority of all jobs waiting in the queue.

Example:

[tkaiser@ra ~]$ mdiag -p
diagnosing job priority information (partition: ALL)

Job                    PRIORITY*   Cred(Accnt:Class)    FS( User:Accnt)  Serv(QTime)
             Weights   --------       1(    1:    1)    20(    5:   21)     1(    2)

64023                     78030     0.0(  0.0:  0.0)   2.4(  5.0: -5.8)  97.6(39978)
78157                     11456     0.0(  0.0:  0.0)  71.7(  2.3: 19.0)  28.3(1622.)
...
...
79029                     10300     0.0(  0.0:  0.0)  79.7(  2.3: 19.0)  20.3(1044.)
79040                     10254     0.0(  0.0:  0.0)  80.1(  2.3: 19.0)  19.9(1021.)
79042                     10236     0.0(  0.0:  0.0)  80.2(  2.3: 19.0)  19.8(1012.)
78138                      8389     0.0(  0.0:  0.0)  53.4( -6.2: 12.2)  46.6(1953.)
78143                      8389     0.0(  0.0:  0.0)  53.4( -6.2: 12.2)  46.6(1953.)
78144                      8389     0.0(  0.0:  0.0)  53.4( -6.2: 12.2)  46.6(1953.)
78145                      8389     0.0(  0.0:  0.0)  53.4( -6.2: 12.2)  46.6(1953.)
78146                      8389     0.0(  0.0:  0.0)  53.4( -6.2: 12.2)  46.6(1953.)
78148                      8363     0.0(  0.0:  0.0)  53.6( -6.2: 12.2)  46.4(1940.)
...
...
...
78384                     -8558     0.0(  0.0:  0.0)  96.8(-26.8:-14.7)   3.2(148.0)
78385                     -8558     0.0(  0.0:  0.0)  96.8(-26.8:-14.7)   3.2(148.0)
79051                     -6103     0.0(  0.0:  0.0)  95.2( -2.5:-14.7)   4.8(162.0)
79052                     -1693     0.0(  0.0:  0.0)  86.8(  4.3: -5.8)  13.2(152.0)

Percent Contribution   --------     0.0(  0.0:  0.0)  55.2( 11.8: 47.8)  44.8( 44.8)

* indicates absolute/relative system prio set on job

[tkaiser@ra ~]$

The first column gives the jobs number and the second the priority. Priorities change over time. A higher number indicates that a greater priority has been assigned to the job scheduler. There are two factors that effect this priority: the time a job has been waiting; and the "Fair Share" algorithm.

The right-hand column contains the time a job has been eligible to run in parentheses (minutes.) Next to this time is a priority related to the wait time.

Under the "FS( User:Accnt)" column there are three numbers. The number outside of the parentheses is again a priority. The inside numbers are the weighting factors for a user and the users account group. If a user has recently made heavy use of RA, then they will have a lower individual weighting. If the rest of the people in the account group have used a lot of time recently, then the second weighting will also be low.