[
https://issues.apache.org/jira/browse/HDDS-15941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chi-Hsuan Huang updated HDDS-15941:
-----------------------------------
Status: Patch Available (was: In Progress)
> Resolve client before block loop and return null from sortDatanodesForWrite
> ---------------------------------------------------------------------------
>
> Key: HDDS-15941
> URL: https://issues.apache.org/jira/browse/HDDS-15941
> Project: Apache Ozone
> Issue Type: Improvement
> Reporter: Chi-Hsuan Huang
> Assignee: Chi-Hsuan Huang
> Priority: Major
> Labels: pull-request-available
>
> Follow\-up to the HDDS\-15059 review feedback in
> [https://github.com/apache/ozone/pull/10633#pullrequestreview\-4746713524:|https://github.com/apache/ozone/pull/10633#pullrequestreview-4746713524:]
> # In OMKeyRequest#allocateBlock, the client is fixed for the whole request,
> but the current code retries client resolution for every pipeline when the
> client cannot be resolved. Resolve the client once before the block loop, and
> when the client cannot be resolved, fall back to SCM\-side sorting by passing
> the remoteAddress to SCM.
> [https://github.com/apache/ozone/pull/10633#discussion\_r3623837258|https://github.com/apache/ozone/pull/10633#discussion_r3623837258]
> # KeyManagerImpl#sortDatanodesForWrite returns the input list instance to
> signal that no sort happened, and the caller checks reference equality
> \(sortedNodes \!= nodes\). This looks like a bug and is fragile \(someone may
> change it to equals later\). Return null instead when no sort happens, and
> have callers check for null.
> [https://github.com/apache/ozone/pull/10633#discussion\_r3623861478|https://github.com/apache/ozone/pull/10633#discussion_r3623861478]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
