[
https://issues.apache.org/jira/browse/HDDS-17302?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Siyao Meng updated HDDS-17302:
------------------------------
Epic Link: HDDS-17321
> OM client retry wait time grows without bound with a single OM
> --------------------------------------------------------------
>
> Key: HDDS-17302
> URL: https://issues.apache.org/jira/browse/HDDS-17302
> Project: Apache Ozone
> Issue Type: Bug
> Components: Ozone Client
> Reporter: Siyao Meng
> Assignee: Siyao Meng
> Priority: Major
> Labels: pull-request-available
>
> With a single OM, when the OM is not reachable (for example right after
> starting a Docker Compose cluster), the OM client waits 2s, 4s, 6s, ...
> between retries, growing by {{ozone.client.wait.between.retries.millis}} each
> time without a cap. With the default
> {{ozone.client.failover.max.attempts=500}}, a command like {{ozone fs -ls
> ofs://om/}} started before OM is ready can keep sleeping long after OM is up.
> Interrupting and rerunning the command works immediately.
> With multiple OMs, the client waits a fixed
> {{ozone.client.wait.between.retries.millis}} after trying all OMs once. With
> a single OM, {{OMFailoverProxyProviderBase#getWaitTime}} instead treats round
> robin failover back to the same OM like a retry requested by the OM (e.g.
> {{OMLeaderNotReadyException}}), which increases the wait linearly.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
