[
https://issues.apache.org/jira/browse/HDDS-7462?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Swaminathan Balachandran updated HDDS-7462:
-------------------------------------------
Description:
EC Reconstruction(with Write Chunk Operation) recreates open with replica Index
0 when StaleRecoveringContainerScrubbingService deletes the recovering
container. Thus an invalid container with replica 0 is created. This could
potentially cause SCM failure when container is reported with heartbeat & also
partial reconstructed container when a new block is written simultaneously with
recovering container being deleted.
Marking the recovering container as unhealthy should fix the issue. Handling
the failure to delete unhealthy container should fix the issue from
Reconstruction Coordinater will cleanup the stale container.
was:EC Reconstruction(with Write Chunk Operation) recreates open with replica
Index 0 when StaleRecoveringContainerScrubbingService deletes the recovering
container. Thus an invalid container with replica 0 is created. This could
potentially cause SCM failure when container is reported with heartbeat & also
partial reconstructed container when a new block is written simultaneously with
recovering container being deleted.
> EC: Fix Reconstruction Issue with StaleRecoveringContainerScrubbingService
> --------------------------------------------------------------------------
>
> Key: HDDS-7462
> URL: https://issues.apache.org/jira/browse/HDDS-7462
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Swaminathan Balachandran
> Assignee: Swaminathan Balachandran
> Priority: Critical
>
> EC Reconstruction(with Write Chunk Operation) recreates open with replica
> Index 0 when StaleRecoveringContainerScrubbingService deletes the recovering
> container. Thus an invalid container with replica 0 is created. This could
> potentially cause SCM failure when container is reported with heartbeat &
> also partial reconstructed container when a new block is written
> simultaneously with recovering container being deleted.
> Marking the recovering container as unhealthy should fix the issue. Handling
> the failure to delete unhealthy container should fix the issue from
> Reconstruction Coordinater will cleanup the stale container.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
