[
https://issues.apache.org/jira/browse/HDDS-3080?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Prashant Pogde updated HDDS-3080:
---------------------------------
Target Version/s: 1.2.0
I am managing the 1.1.0 release and we currently have more than 600 issues
targeted for 1.1.0. I am moving the target field to 1.2.0.
If you are actively working on this jira and believe this should be targeted to
1.1.0 release, Please change the target field back to 1.1.0 before Feb 05,
2021.
> Configurable container placement policy may not be compatible with pipeline
> placement policy
> --------------------------------------------------------------------------------------------
>
> Key: HDDS-3080
> URL: https://issues.apache.org/jira/browse/HDDS-3080
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Components: SCM
> Affects Versions: 1.0.0
> Reporter: Stephen O'Donnell
> Priority: Major
>
> New pipelines are created on hosts in a rack aware way using
> PipelinePlacementPolicy, which supports only rack aware and random.
> However the Container Placement Policy used by ReplicationManager can be
> configured between the following using the config option
> ozone.scm.container.placement.impl:
> SCMContainerPlacementRackAware
> SCMContainerPlacementCapacity
> SCMContainerPlacementRandom
> If there is a network topology configured via net.topology.table.file.name,
> then it would be possible to have an incompatible setting for
> ozone.scm.container.placement.impl.
> We need to consider how to ensure the replication manager and Pipeline
> Provider always use compatible policies. One option may be to merge the rack
> aware policies as suggested in HDDS-3079, and also allow Capacity and Random
> to be used in the pipeline provider.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
