refreshLinks.php --dfn-only cron jobs do not seem to be running
Closed, ResolvedPublic
Actions

Assigned To

None

Authored By

	PleaseStand
	May 3 2015, 9:00 AM

Description

The monthly refreshLinks.php --dfn-only cron jobs do not seem to be running. A straightforward SQL query returns rows that should have been deleted, and a check of the deletion logs for those page IDs suggests that this has been the case since at least August 2014.

Perhaps someone could check /var/log/mediawiki/refreshLinks/s2@2.log on terbium, or some other relevant log file, for an indication of an error?

On that particular wiki, all the pages were in the File namespace, which suggests that there may be a bug in MediaWiki. That, however, would be a separate bug from any that might be affecting the cron jobs.

Related Objects
Search...

Status	Assigned	Task
Resolved	None	T25816 Missing file (red link) tracking category support
Resolved	None	T26480 Special:WantedCategories on Italian Wikiquote is showing two orphaned categories since years ago
Resolved	Krenair	T18112 Run "refreshLinks.php --dfn-only" on all wikis periodically
Resolved	None	T97926 refreshLinks.php --dfn-only cron jobs do not seem to be running

Event Timeline

PleaseStand created this task.May 3 2015, 9:00 AM

PleaseStand raised the priority of this task from to Needs Triage.

PleaseStand updated the task description. (Show Details)

PleaseStand added a project: Wikimedia-Site-requests.

PleaseStand subscribed.

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptMay 3 2015, 9:00 AM

I don't know about eowiki, but enwiki has quite a queue for refreshLinks:

krenair@terbium:~$ mwscript showJobs.php enwiki --group
refreshLinks: 6829959 queued; 21 claimed (0 active, 21 abandoned); 0 delayed

It was much worse than that for several days, and I mentioned it on IRC...

krenair@terbium:~$ ls -al /var/log/mediawiki/refreshLinks/s*
-rw-rw-r-- 1 mwdeploy mwdeploy   9928 Jun  2  2014 /var/log/mediawiki/refreshLinks/s2@2.log
-rw-rw-r-- 1 mwdeploy mwdeploy 496748 Jun  3  2014 /var/log/mediawiki/refreshLinks/s3@3.log
-rw-rw-r-- 1 www-data www-data    579 Apr  4 02:14 /var/log/mediawiki/refreshLinks/s4@4.log
-rw-rw-r-- 1 www-data www-data   1177 Apr  5 00:56 /var/log/mediawiki/refreshLinks/s5@5.log
-rw-rw-r-- 1 www-data www-data   1782 Apr  6 01:04 /var/log/mediawiki/refreshLinks/s6@6.log
-rw-rw-r-- 1 www-data www-data   6920 Apr  7 01:36 /var/log/mediawiki/refreshLinks/s7@7.log

Krenair added a project: acl*sre-team.May 3 2015, 4:10 PM

https://github.com/wikimedia/operations-puppet/blob/98d7853ce884edd5478d31a72369255b8c7b4f6f/manifests/misc/maintenance.pp#L12

s1 doesn't have a refreshLinks job

And yuvi just fixed the permissions on the s2 and s3 logs so hopefully they should start getting written to again

Today, the script did run on dewiki (in s5); see the query results linked from T98110.

I think the differing file permissions did prevent the script from running on s2 and s3 wikis. We shall see...

[terbium:/var/log/mediawiki/refreshLinks] $ ls
total 748K
4.0K drwxrwxr-x 2 www-data mwdeploy 4.0K Jun  7  2014 .
4.0K drwxr-xr-x 5 www-data wikidev  4.0K Mar 29 05:00 ..
 12K -rw-rw-r-- 1 www-data www-data 9.7K Jun  2  2014 s2@2.log
492K -rw-rw-r-- 1 www-data www-data 486K Jun  3  2014 s3@3.log
108K -rw-rw-r-- 1 www-data www-data 102K May  4 01:59 s4@4.log
 72K -rw-rw-r-- 1 www-data www-data  66K May  5 01:15 s5@5.log
 44K -rw-rw-r-- 1 www-data www-data  41K May  6 00:44 s6@6.log
 12K -rw-rw-r-- 1 www-data www-data 8.2K May  7 00:11 s7@7.log

I ran the update command for s2 manually and it appeared to run just normal.

In T97926#1255734, @Reedy wrote:

https://github.com/wikimedia/operations-puppet/blob/98d7853ce884edd5478d31a72369255b8c7b4f6f/manifests/misc/maintenance.pp#L12

s1 doesn't have a refreshLinks job

That's T18112: Run "refreshLinks.php --dfn-only" on all wikis periodically, allegedly still blocked on T44180: make "refreshLinks.php --dfn-only" faster (to work on en.wiki).

ArielGlenn triaged this task as Medium priority.May 12 2015, 1:50 PM

ArielGlenn set Security to None.