This message was deleted Apache Druid #troubleshooting

This message was deleted.

Slackbot

11/28/2022, 5:49 PM

This message was deleted.

Sergio Ferragut

11/28/2022, 8:03 PM

Segment replicas on other Historicals will help.

Karan Kumar

11/29/2022, 4:46 AM

A query retry would be the only way I guess

Gian Merlino

11/30/2022, 8:23 AM

if you can kill the pod cleanly (i.e. sigterm the proc inside it, and wait for it to exit) then you will see fewer failures

Gian Merlino

11/30/2022, 8:23 AM

as during a clean exit, the historical waits for currently-running queries to finish before shutting down