test JAWS nmdc perlmutter site

[aclum]

The JAWS team has set up a site on Perlmutter for NMDC. Test that this works on an example workflow.

[aclum]

Slacked Elais as this still appears to be trying to use cori cscratch which doesn't exist anymore.

[ssarrafan]

Alicia is in the midst of JAWS testing. Will move to Zoe sprint.

[aclum]

The jaws team is working on some perlmutter issues that impacted my test submission. Dani will bring this up at the JAWS Monday meeting

jaws log 65101

STATUS_FROM STATUS_TO TIMESTAMP COMMENT

created upload queued 2023-06-28 10:28:02
upload queued upload complete 2023-06-28 10:28:02
upload complete submission failed 2023-06-28 10:28:14 RPC function error: Instance <Run at 0x150a915acbb0> is not bound to a Session; attribute refresh operation cannot proceed (Background on this error at: https://sqlalche.me/e/20/bhk3)
submission failed email sent 2023-06-28 10:28:24
email sent done 2023-06-28 10:28:35

[aclum]

This is blocked on a new JAWS release which is expected next week, needs to be moved to the next sprint.

[ssarrafan]

JAWS release is still pending, they are reviewing some errors. Will move this to the next sprint. @aclum let me know if you prefer for this to go to the backlog till after ESA.

[aclum]

Yes, this should be in the backlog until after ESA.

[aclum]

Still no ETA on a release date from the JAWS team.

[aclum]

The new version JAWS was released last week. I ran a test this morning and it failed and I was unable to debug b/c of permission issues so filed a ticket for that. https://code.jgi.doe.gov/advanced-analysis/jaws-support/-/issues/111

[aclum]

The perlmutter JAWS site is working with cds_prediction.wdl. Current blockers 1) when the workflow prefix matches the site name 2) issue with runtime_minutes

[aclum]

backlogging this issue until blockers are resolved.

[aclum]

workflow prefix matches the site name is fixed but not released yet.

[aclum]

prefix issue fixed was released by the jaws team. I ran an issue yesterday and there was an issue with /refdata mount. See https://code.jgi.doe.gov/advanced-analysis/jaws-site/-/issues/1756 This is fixed but I don't believe deployed yet.

[aclum]

A filtering test run on Dec 18 2023 completed successfully so the volume mounting appears to be working. Perlmutter is down today.

[aclum]

Closing as I believe all issues have been resolved. Will open a new ticket if new issues come up.