sijie commented on issue #1625: Bookie's AutoRecovery affects publish times URL: https://github.com/apache/incubator-pulsar/issues/1625#issuecomment-384835295 @bardock > if a single bookie goes down does that mean bookie machine physically gone? or just (docker) process being killed. if a single bookie goes down is temporary, you can try to set `lostBookieRecoveryDelay` so the auto recovery will kick in later. That would avoid unneeded replication. if a single bookie goes down is permanently and you don't want to throttle the replication progress, you probably can move autorecovery to detailed hosts. > Run autorecovery process on dedicated nodes in our BookKeeper cluster. I believe it is not supported in bk version shipped with pulsar 1.20. Is it possible to use bk 4.7 only for autorecovery nodes? you actually can run `bin/bookkeeper autorecovery` with pulsar package. You can check this k8s deployment yaml file - https://github.com/apache/incubator-pulsar/blob/master/deployment/kubernetes/google-kubernetes-engine/bookie.yaml#L183 this is how it runs bookie autorecovery in separated pods. Hope this help with your situation.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
