Fairly regularly I'm seeing the following message showing up in /var/log/koha/instance/es-indexer-output.log [2024/05/20 09:28:13] [WARN] Update of elastic index failed with: Exception 'Koha::Exceptions::Elasticsearch::BadResponse' thrown 'Bad response received when submitting request to Elasticsearch' with details => [https://hostname:9300]-[599] Timed out while waiting for socket to become ready for reading, type => Timeout main::catch {...} /usr/share/koha/bin/workers/es_indexer_daemon.pl (229) But this doesn't really tell me anything useful. It's not clear what records failed to update or what the job number was, so there's no way to re-try and fix the missing indexed data. Also, I've updated the request_timeout from 30 to 60, so it's a fairly long timeout. Generally speaking, there was a fair amount of indexing activity, but around 09:27:13 there wasn't. Although there was 6 jobs at 09:28:13, so maybe it's not the request_timeout that needs tweaking...
"Timed out while waiting for socket to become ready for reading" comes from HTTP::Tiny, which is the HTTP client used by Search::Elasticsearch. The only direct mention I've found for it is on the Elastic discussion forms for "Elasticsearch - Error Timeout 599". (Not including the link so that Bugzilla doesn't block me.)
Actually, I forgot I'd also found a mention on the elasticsearch-perl Github for issue "reindex timeout #112"
It sounds like the bug description is a bit misleading/sounds too harmless, isn't the real problem here that we are missing records to index?
Yeah, I think I was barking up the wrong tree here...