Addshore added a comment.
I chatted a little with @Krinkle and @daniel and ...
Firstly:
- The `purgeWebCache` correctly triggers a `UpdateHtmlCacheJob` which runs
quickly and "dirties" the parsercache for the page being dispatched to
- This leads to cache misses when using a debug line like
`(MediaWiki\MediaWikiServices::getInstance()->getParserCache())->get(Title::newFromText('Category:Interior
of Chapelle des Anges de Neunhoffen à Dambach
(Bas-Rhin)')->toPageRecord(),ParserOptions::newCanonical())->getCacheTime();`
- A user view will parse the page correctly, and the page property for
wikibase_item will be visible in the parser output with
`(MediaWiki\MediaWikiServices::getInstance()->getParserCache())->get(Title::newFromText('Category:Interior
of Chapelle des Anges de Neunhoffen à Dambach
(Bas-Rhin)')->toPageRecord(),ParserOptions::newCanonical())->getProperty('wikibase_item');`
- At this stage the page_prop will not be in the page_props table.
Secondly:
- The `scheduleRefreshLinks` should trigger a `RefreshLinksJob`. (These take
a little longer to run in most cases)
- At some unknown time in the future the `page_props` table should be
populated with the `wikibase_item` page property. (we saw this happen a few
times)
- You'd be able to see this in the DB, or using something like
`PageProps::getInstance()->getProperties([Title::newFromText('Category:Interior
of Chapelle des Anges de Neunhoffen à Dambach (Bas-Rhin)')],'wikibase_item');`
-----
Now trying to apply this again to the random failures seen in T233520:
page_props wikibase_item is sometimes not added to client pages when a sitelink
is added on a repo <https://phabricator.wikimedia.org/T233520> where
`page_props` is seemingly never populated....
One can make a WikiPageUpdater for a client with:
$services = MediaWiki\MediaWikiServices::getInstance();
$pu = new \Wikibase\Client\Changes\WikiPageUpdater(
JobQueueGroup::singleton(), \Wikibase\Client\WikibaseClient::getLogger(
$services ), $services->getStatsdDataFactory());
And then trigger a web cache job with something like:
$pu->purgeWebCache( [ Title::newFromText('Category:Interior of Chapelle des
Anges de Neunhoffen à Dambach (Bas-Rhin)') ],
['rootJobTimestamp'=>wfTimestampNow()], 'update', 'uid:addshore' );
As with the bullet point lists above, we see the cache start missing at this
point, and a web view will re parse the page.
When trying to do the same with the page for the refreshlinks job, I can do
the following:
$pu->scheduleRefreshLinks( [ Title::newFromText('Category:Interior of
Chapelle des Anges de Neunhoffen à Dambach (Bas-Rhin)') ],
['rootJobTimestamp'=>wfTimestampNow()], 'update', 'uid:addshore' );
In theory this should schedule the job, and no error occurs.
But in both of these job cases I do not see anything matching the job I
scheduled in kafka (perhaps I'm looking in the wrong place).
I waited 10-30 minutes and did not see the `page_props` for a category page
on commons end up updating. (could be the job was still in the queue?)
I then manually ran the job on the debug server:
(new RefreshLinksJob( Title::newFromText('Category:Interior of Chapelle des
Anges de Neunhoffen à Dambach (Bas-Rhin)'),
['rootJobTimestamp'=>wfTimestampNow()] ))->run();
And the page_prop now instantly appears in the page_props table
>>>
PageProps::getInstance()->getProperties([Title::newFromText('Category:Interior
of Chapelle des Anges de Neunhoffen à Dambach (Bas-Rhin)')],'wikibase_item');
=> [
103128055 => "Q106684122",
]
---------
So either:
1. The call to scheduleRefreshLinks is not actually scheduling a job? and
thus the page_props table is never being updated?
2. The jobs for the pages that appear to not have page_props just have not
executed yet?
3. something else?
Looking at
https://grafana.wikimedia.org/d/CbmStnlGk/jobqueue-job?orgId=1&var-dc=eqiad%20prometheus%2Fk8s&var-job=refreshLinks
though these jobs shoul only take a few hours at most?
But really this is ~15 mins currently.
This lines up with what @Michael @Jakob_WMDE and I were seeing between
testwiki and testwikidata, but not with what is seen in T233520
<https://phabricator.wikimedia.org/T233520> on commonswiki?
TASK DETAIL
https://phabricator.wikimedia.org/T280627
EMAIL PREFERENCES
https://phabricator.wikimedia.org/settings/panel/emailpreferences/
To: Michael, Addshore
Cc: Krinkle, daniel, Jakob_WMDE, Mike_Peel, Aklapper, Jheald,
Lucas_Werkmeister_WMDE, Addshore, WMDE-leszek, aaron, tstarling, LibrErli,
Lea_Lacroix_WMDE, Ladsgroup, RolandUnger, Urbanecm, Bencemac, Tacsipacsi,
Kizule, CCicalese_WMF, Lydia_Pintscher, Invadibot, maantietaja, Muchiri124,
Hazizibinmahdi, CBogen, Akuckartz, Iflorez, WDoranWMF, alaa_wmde, holger.knust,
EvanProdromou, Nandana, Lahi, Gq86, Ramsey-WMF, GoranSMilovanovic, QZanden,
LawExplorer, Poyekhali, _jensen, rosalieper, Agabi10, Taiwania_Justo,
Scott_WUaS, Pchelolo, Jonas, Ixocactus, Wong128hk, Wikidata-bugs, aude,
El_Grafo, Dinoguy1000, Steinsplitter, Mbch331, Keegan
_______________________________________________
Wikidata-bugs mailing list -- wikidata-bugs@lists.wikimedia.org
To unsubscribe send an email to wikidata-bugs-le...@lists.wikimedia.org
