Our cirrussearch dumps have all failed with the same error at the start of the weekly dump.
https://airflow-test-k8s.wikimedia.org/dags/mediawiki_cirrussearch_dump/grid
Note that the dumps are also still running on snapshot1016, although these are legacy dumps and are about to be switched off.
The logs of the failed jobs all say this:
[2025-07-21, 17:30:34 UTC] {pod_manager.py:520} INFO - [base] Named cluster (dnsdisc) is not configured for maintenance operations. Allowed clusters: eqiad, codfw, cloudelastic
[2025-07-21, 17:30:34 UTC] {pod_manager.py:520} INFO - [base] extensions/CirrusSearch/maintenance/DumpIndex.php failed for /mnt/dumpsdata/otherdumps/cirrussearch/20250721/enwiki-20250721-cirrussearch-content.json.gzHere is an example.




