Page MenuHomePhabricator

Toolforge grid jobs stuck with: can't get password entry for user "tools.<tool>". Either user does not exist or error with NIS/LDAP etc.
Closed, ResolvedPublic

Description

I've had multiple grid jobs get stuck with an error entry of can't get password entry for user "tools.<tool>". Either user does not exist or error with NIS/LDAP etc.

In each case qstat shows the job in Eqw state, which I guess means that jsub -once will not run a new job.

Event Timeline

@Majavah believes this is caused by stuff like T282474: tools.topicmatcher update_items_from_sparql.php frequently running Toolforge nodes out of resources which is causing the node to run out of resources, so new jobs will somehow have issues talking to LDAP.

taavi claimed this task.

Haven't seen this in a while. Closing.