From just looking at https://gerrit.wikimedia.org/r/plugins/gitiles/phabricator/antivandalism/+/refs/heads/wmf/stable/scripts/init/init-script.php , that code must be broken since https://secure.phabricator.com/rPHU720c8116845bb9dc19334170e6c0702aa0210c78 in February 2020.
Description
Details
Related Objects
Event Timeline
Indeed running php ./libext/ava/scripts/score.php says triggers that error. After making code changes, output is expected man page style.
I still remain unsuccessful in printing a score though when passing a username to php ./score.php score --user username.
Change 933907 had a related patch set uploaded (by Aklapper; author: Aklapper):
[phabricator/antivandalism@wmf/stable] AVA: Make score.php not fail with Fatal Error after libphutil removal
@Aklapper did you run the php ./libext/ava/scripts/score.php round 15:30, so 8 minutes ago. There was a large spike in latenacy on Phabricator around there.
@SLyngshede-WMF I did not run this on any production server (and I think I cannot anyway due to lack of full shell access), only locally.
This code should not trigger much load anyway, as it would be calculating a score for one single user.
I can see that spike at https://grafana.wikimedia.org/d/000000587/phabricator . Not sure where it originated from; I don't see anything suspicious in https://phabricator.wikimedia.org/daemon/log/ or https://phabricator.wikimedia.org/feed/ or https://phabricator.wikimedia.org/herald/transcript/ .
I have been mass-editing the View Policy of ~800 spam tasks via API calls an hour ago between 12:35 and 13:14UTC though.
For the records, on 2023-06-17 I re-enabled H293 (applied to every and any task activity) in non-public T338611#8944935 so I generally expect a higher load.
@Aklapper Thank you for looking, I didn't really see anything suspicious in the logs either.
Change 933907 merged by Brennen Bearnes:
[phabricator/antivandalism@wmf/stable] AVA: Make score.php not fail with Fatal Error after libphutil removal