Administration

Showing 45

brequeue on zombie jobs please

a machine crashed on an on-premise cluster leaving many jobs in a UNKWN state. as described here [https://www-01.ibm.com/support/docview.wss?uid=isg3T1027518] i tried bkill -r twice then brequeue -e, but got the error message "Job has already fini...

over 4 years ago in Spectrum LSF / Administration 1 Not under consideration

2 VOTE

bhosts -o "non shared dynamic numeric elim resource"

I'd like to see "non shared dynamic numeric elim resource" info from bhosts -o. bhosts -o can only view LSF embedded resources.(as previously defined)"bhosts -o "field_name[:[-][output_width]]" lsload -o could tell us non-shared resource status. B...

almost 5 years ago in Spectrum LSF / Administration 0 Not under consideration

1 VOTE

Provide Minor Enhancement to 'bjobs -env jobid' Command to search and return variables server side

The 'bjobs -env jobid' command is very useful, but in our environment, the job environment can be very large. We plan to use this command as a part of our ESUB modify controls and it could be called quite frequently. Therefore, it will generate a ...

over 2 years ago in Spectrum LSF / Administration 3 Not under consideration

2 VOTE

blimits -n to select multiple limit name

blimits man pages says that's possible to select multiple limit names but it's not functional. can you please implement this functionality ?

over 5 years ago in Spectrum LSF / Administration 0 Not under consideration

1 VOTE

Support bimages for podman

Hello, we want to enable users running podman to perform management of images downloaded by the podman. The management includes clean-up of images based on some policy like LRU or any other that user decides. I think bimages is supported for docke...

almost 3 years ago in Spectrum LSF / Administration 2 Not under consideration

1 VOTE

Make lsadmin/badmin accept "yes" and "no" as answers

Questions asked by `lsadmin` and `badmin` (and maybe other administrative LSF commands) only accept "y" and "n" as answers. "yes" and "no" are not accepted but should be

about 3 years ago in Spectrum LSF / Administration 0 Not under consideration

1 VOTE

bmod jobs for per queue, or per user

Currently bmod runs per job so if there are 20000 running jobs, I've to run bmod 20000 times, one for each job. Can bmod have functionality like bkill to update bulk jobs , based on certain criteria like for user, queue, hostgroup etc ?

over 3 years ago in Spectrum LSF / Administration 3 Not under consideration

1 VOTE

How to start/stop/restart Suite services on all the nodes in Suite cluster

Currently cluster admin needs to run "systemctl" for all the suite services. It's hard to remember which Suite services are running on which role of nodes in Suite cluster - different services are on master, GUI and compute nodes. So there is a ne...

almost 4 years ago in Spectrum LSF / Administration 0 Not under consideration

1 VOTE

candidate host pending reason

It's about pending reason hostname. Current pending reason as below. These 3 hosts are not in host group of this queue, but they are in candidate host list. PENDING REASONS: Candidate host pending reasons (18 of 7574 hosts): Job slot limit reached...

over 4 years ago in Spectrum LSF / Administration 0 Not under consideration

1 VOTE

Allow `lsadmin reconfig`to run in batch mode for only the master candidate hosts

`lsadmin reconfig` cannot be run in unattended mode only on master and master candidate hosts. It would be nice to have such an option. `lsadmin limrestart pf host1 host2` is not an acceptable replacement for this functionality. For details see "U...

over 4 years ago in Spectrum LSF / Administration 0 Not under consideration

Please enter your email address

FILTER BY CATEGORY