Steps to replicate the issue (include links if applicable):
What happens?:
What should have happened instead?:
The proper count should be displayed.
Other information (browser name/version, screenshots, etc.):
Might be related to T400908
On the first day of each month the maintenance script recountCategories.php --mode all is running to correct some wrong category counts. It seems today it broke the counts.
Maybe from work on T299951: Normalize categorylinks table?
Change #1175089 had a related patch set uploaded (by Ladsgroup; author: Amir Sarabadani):
[mediawiki/core@master] recountCategories: Avoid escpaing column name
Change #1175095 had a related patch set uploaded (by Ladsgroup; author: Amir Sarabadani):
[mediawiki/core@wmf/1.45.0-wmf.12] recountCategories: Avoid escpaing column name
Change #1175095 merged by jenkins-bot:
[mediawiki/core@wmf/1.45.0-wmf.12] recountCategories: Avoid escpaing column name
Change #1175089 merged by jenkins-bot:
[mediawiki/core@master] recountCategories: Avoid escpaing column name
Mentioned in SAL (#wikimedia-operations) [2025-08-01T12:46:22Z] <ladsgroup@deploy1003> Started scap sync-world: Backport for [[gerrit:1175095|recountCategories: Avoid escpaing column name (T400987)]]
Mentioned in SAL (#wikimedia-operations) [2025-08-01T12:48:30Z] <ladsgroup@deploy1003> ladsgroup: Backport for [[gerrit:1175095|recountCategories: Avoid escpaing column name (T400987)]] synced to the testservers (see https://wikitech.wikimedia.org/wiki/Mwdebug). Changes can now be verified there.
Mentioned in SAL (#wikimedia-operations) [2025-08-01T12:54:58Z] <ladsgroup@deploy1003> Finished scap sync-world: Backport for [[gerrit:1175095|recountCategories: Avoid escpaing column name (T400987)]] (duration: 08m 36s)
Mentioned in SAL (#wikimedia-operations) [2025-08-01T12:57:37Z] <Amir1> re-running recountCategories.php on all wikis except s4 and s1 (T400987)
Well, it caused a massive replag in s7 so I had top stop it. It's fully done in a lot of sections.
I restarted it on s2 and s3 (the rest were not needed or were done). Only thing stopped right now is s7.
Only s3 and s7 left and thankfully I found the --throttle option and added a 1s sleep between each batch. That should slow it down.
We have the same problem in the Norwegian Bokmål Wikipedia.
See https://no.wikipedia.org/wiki/Kategori:Sider_som_bruker_Kartographer-utvidelsen
Subcatgories are reported as empty, and there's no total count of pages in the category, it just says "the following 200 pages are in this category".
Edit: I didn't read Amir's comments thoroughly enough – the fix is being applied to all wikis, so I guess my added report is superfluous. Thanks for fixing!
It was broken halfway through nlwiki. Now it's updating plwiki. nowiki should be fine now.
I think it also happened in ukwiki and ruwiki. See https://uk.wikipedia.org/wiki/%D0%9A%D0%B0%D1%82%D0%B5%D0%B3%D0%BE%D1%80%D1%96%D1%8F:%D0%A1%D0%BF%D0%BE%D1%80%D1%82_%D1%83_%D0%9D%D1%96%D0%BC%D0%B5%D1%87%D1%87%D0%B8%D0%BD%D1%96_%D0%B7%D0%B0_%D1%80%D0%BE%D0%BA%D0%BE%D0%BC https://ru.wikipedia.org/wiki/%D0%9A%D0%B0%D1%82%D0%B5%D0%B3%D0%BE%D1%80%D0%B8%D1%8F:%D0%9D%D0%B0%D0%B3%D1%80%D0%B0%D0%B6%D0%B4%D1%91%D0%BD%D0%BD%D1%8B%D0%B5_%D0%BC%D0%B5%D0%B4%D0%B0%D0%BB%D1%8F%D0%BC%D0%B8_%D0%A1%D0%BB%D0%BE%D0%B2%D0%B0%D0%BA%D0%B8%D0%B8
@Ladsgroup Our counts. on Serbian Wikipedia, are still broken. So fix was not applied everywhere?
I think it didn't run fully because of k8s issues. I run it again everywhere with the new method rzl suggested.
ladsgroup@deploy1003:~$ mwscript-k8s --dblist s2 -- recountCategories.php --mode=all --throttle 1000 ⏳ Starting recountCategories.php on Kubernetes as job mw-script.eqiad.xvc10k65 ... 🚀 Job is running. For streaming logs, run: K8S_CLUSTER=eqiad KUBECONFIG=/etc/kubernetes/mw-script-eqiad.config kubectl logs -f job/mw-script.eqiad.xvc10k65 mediawiki-xvc10k65-app ladsgroup@deploy1003:~$ mwscript-k8s --dblist s3 -- recountCategories.php --mode=all --throttle 1000 ⏳ Starting recountCategories.php on Kubernetes as job mw-script.eqiad.qmfca8me ... 🚀 Job is running. For streaming logs, run: K8S_CLUSTER=eqiad KUBECONFIG=/etc/kubernetes/mw-script-eqiad.config kubectl logs -f job/mw-script.eqiad.qmfca8me mediawiki-qmfca8me-app ladsgroup@deploy1003:~$ mwscript-k8s --dblist s5 -- recountCategories.php --mode=all --throttle 1000 ⏳ Starting recountCategories.php on Kubernetes as job mw-script.eqiad.hqlge2r3 ... 🚀 Job is running. For streaming logs, run: K8S_CLUSTER=eqiad KUBECONFIG=/etc/kubernetes/mw-script-eqiad.config kubectl logs -f job/mw-script.eqiad.hqlge2r3 mediawiki-hqlge2r3-app ladsgroup@deploy1003:~$ mwscript-k8s --dblist s6 -- recountCategories.php --mode=all --throttle 1000 ⏳ Starting recountCategories.php on Kubernetes as job mw-script.eqiad.3ry0rlyb ... 🚀 Job is running. For streaming logs, run: K8S_CLUSTER=eqiad KUBECONFIG=/etc/kubernetes/mw-script-eqiad.config kubectl logs -f job/mw-script.eqiad.3ry0rlyb mediawiki-3ry0rlyb-app ladsgroup@deploy1003:~$ mwscript-k8s --dblist s7 -- recountCategories.php --mode=all --throttle 1000 ⏳ Starting recountCategories.php on Kubernetes as job mw-script.eqiad.2qqdtcw1 ... 🚀 Job is running. For streaming logs, run: K8S_CLUSTER=eqiad KUBECONFIG=/etc/kubernetes/mw-script-eqiad.config kubectl logs -f job/mw-script.eqiad.2qqdtcw1 mediawiki-2qqdtcw1-app
FTR