Page MenuHomePhabricator

Determine cause/fix cloudelastic cross-cluster search missing index errors
Closed, ResolvedPublic

Description

While investigating T355795 ,@EBernhardson noticed noticed that a number of errors in Logstash related to missing indices .

Creating this ticket to investigate these cross-cluster search missing index errors (and hopefully fix them).

Event Timeline

bking renamed this task from Determine cause/fix cross-cluster search missing index errors to Determine cause/fix cloudelastic cross-cluster search missing index errors.Mar 4 2024, 4:42 PM
Gehel claimed this task.

This is related to the same cross cluster setting configuration issue.

Reopening, as we could fix this for good by rendering the YAML file in this puppet plan as json and writing it to the cluster settings periodically (such as with a systemd timer).

bking reopened this task as In Progress.Apr 9 2024, 9:48 PM
bking claimed this task.
bking lowered the priority of this task from High to Medium.

Change #1018360 had a related patch set uploaded (by Bking; author: Bking):

[operations/puppet@production] WIP: elasticsearch: prevent cross cluster seed config drift

https://gerrit.wikimedia.org/r/1018360

Change #1018360 merged by Bking:

[operations/puppet@production] elasticsearch: prevent cross cluster seed config drift

https://gerrit.wikimedia.org/r/1018360

Change #1020326 had a related patch set uploaded (by Bking; author: Bking):

[operations/puppet@production] Elastic: unique systemd timer name per cluster

https://gerrit.wikimedia.org/r/1020326

Change #1020326 merged by Bking:

[operations/puppet@production] Elastic: unique systemd timer name per cluster

https://gerrit.wikimedia.org/r/1020326

We've merged the above patches, and everything looks OK so far (timers are active, Elastic API is accepting the PUT requests, the cluster configuration looks as we expect). As such, I'm closing out this ticket...