Change #1104598 had a related patch set uploaded (by DCausse; author: DCausse):

[operations/mediawiki-config@master] cirrussearch: increase shard count for cebwiki_content

https://gerrit.wikimedia.org/r/1104598

Change #1104598 merged by jenkins-bot:

[operations/mediawiki-config@master] cirrussearch: increase shard count for cebwiki_content

https://gerrit.wikimedia.org/r/1104598

Mentioned in SAL (#wikimedia-operations) [2024-12-17T14:05:33Z] <dcausse@deploy2002> Started scap sync-world: Backport for [[gerrit:1099727|rdf-streaming-updater: add wdqs udpater streams in event stream config (T374919)]], [[gerrit:1104598|cirrussearch: increase shard count for cebwiki_content (T379002)]]

Mentioned in SAL (#wikimedia-operations) [2024-12-17T14:14:55Z] <dcausse@deploy2002> dcausse: Backport for [[gerrit:1099727|rdf-streaming-updater: add wdqs udpater streams in event stream config (T374919)]], [[gerrit:1104598|cirrussearch: increase shard count for cebwiki_content (T379002)]] synced to the testservers (https://wikitech.wikimedia.org/wiki/Mwdebug)

Mentioned in SAL (#wikimedia-operations) [2024-12-17T14:25:41Z] <dcausse@deploy2002> Finished scap sync-world: Backport for [[gerrit:1099727|rdf-streaming-updater: add wdqs udpater streams in event stream config (T374919)]], [[gerrit:1104598|cirrussearch: increase shard count for cebwiki_content (T379002)]] (duration: 20m 07s)

Mentioned in SAL (#wikimedia-operations) [2025-03-06T19:09:02Z] <ebernhardson> T379002 start reindex of cirrus cebwiki_content index in eqiad

Mentioned in SAL (#wikimedia-operations) [2025-03-06T19:11:27Z] <ebernhardson> T379002 start reindex of cirrus cebwiki_content index in codfw

cloudelastic reindex failed to start:

+----------+---------------------------+--------------------------------------+
| error    | rawmessage                | wgCirrusSearchOptimizeIndexForExperi |
|          |                           | mentalHighlighter is set to true but |
|          |                           |  the 'experimental-highligh...       |
+----------+---------------------------+--------------------------------------+

This is likely to be a mixed cluster problem since we renamed experimental-highlighter to cirrus-highlighter for opensearch. Specifically when checking what plugins exist in the cluster we go through each node and only report the plugins that are the installed on all nodes in the cluster. The alias we apply to allow to work with either name comes later into the process. We could update the code to handle this, but should we? It's not clear we should be doing much to encourage running in a mixed cluster mode, but maybe it would be useful for this to still work in a mixed cluster mode.

eqiad and codfw have finished reindexing. We can't consider this done until cloudelastic has been reindexed as well though. That has to wait until it has finished migrating to opensearch 1.3 (and will i suppose be at least a small test of the reindexing automation under opensearch)

Mentioned in SAL (#wikimedia-operations) [2025-03-24T17:51:01Z] <ebernhardson> T379002 Start reindex of cebwiki search indices in cloudelastic

cloudelastic migration to opensearch finished, the reindex is now running. This is also a good verification that our reindexing process still works as expected on opensearch.