Skip to content

Releases: CodyKank/node_search

v1.2.1: Edit module loading config in node_search.sh

09 Jul 14:27
Compare
Choose a tag to compare
Instead of always loading a python3 module which may spit
annoying messages to the screen, a check is performed if python3
is already in the path. If so, a module is not loaded.

v1.2.0: Fixed incorrect counting of free cores in host --details

05 Jul 20:21
Compare
Choose a tag to compare

Better handling of memory output from Xymon tool. Added capability to search the entire long queue and only count the hostgroups a user has access to through the usersets. -uf now displays hostgroups from the long queue and other queues, whereas before it would only search the long queue.

Memory Usage

15 Sep 18:53
Compare
Choose a tag to compare

All issues and bugs up to this point have been patched, specifically in 350ae55, ee3e567, and df774cc. Now a user has the ability to use node_search to accurately monitor memory usage, from not only the standpoint of total usage on a node, but the specific usage of all processes spawned from a user's job. UGE cannot see processes spawned from jobs so node_search picks up the slack allowing a user to track those processes through the --details flag.

v1.0.4

08 Sep 19:55
Compare
Choose a tag to compare

Update for the --details flag for a user. See commit 7c1f1c6.

v1.0.3: Added --all capability. Cleaned up user's --details output.

18 May 20:30
Compare
Choose a tag to compare
The --all capability works now, going through and finding each node without duplication due to interconnection of hostgroups. The output of a user's --details is now less cluttered and more appealing. It displays all jobs running on nodes the specified user's jobs are running on. Also display's virtual memory. Slightly changed the amount of characters the screenwidth is when using --visual or -v, from 120 to 119 and only displaying 39 cores on a line instead of 40. This is done as when using split panes in Tmux it looked funky, so to have it line up better it uses one less core per line.