jojochuang commented on PR #11682: URL: https://github.com/apache/ozone/pull/11682#issuecomment-6477411994
Optional, non-blocking documentation suggestion: we could update the [HA Client Failover documentation](https://ozone.apache.org/docs/next/administrator-guide/configuration/high-availability/client-failover/) to distinguish the two retry-delay cases. This does not need to hold up this PR and can be done later in a separate task. The page currently describes a fixed 2-second delay between every failover retry, but the policy distinguishes: - **Round-robin failover:** wait for `ozone.client.wait.between.retries.millis` after each complete round. With this fix, a single-OM connection failure completes a round on every retry and therefore uses the fixed delay. - **Explicit same-OM retry**, such as `OMLeaderNotReadyException`: increase the delay linearly—2s, 4s, 6s, etc. with defaults. There is no separate delay cap; retries are bounded by `ozone.client.failover.max.attempts`. It would also help to mention that the fix reduces accumulated sleep for 500 single-OM connection-failure retries from approximately 70 hours to 17 minutes, excluding connection/RPC time. No configuration defaults change. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
