[
https://issues.apache.org/jira/browse/YARN-710?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13663536#comment-13663536
]
Siddharth Seth commented on YARN-710:
-------------------------------------
bq. Sid, what would be 'relevant' Yarn records? I would say, this should be
avail for all records. I'm OK moving the write()/read() methods to another
class, RecordIOUtils
Should a YARN utility provide serialization for records like
StartContainerRequest, AllocateResponse (temporary data) ?
bq. Mmhhh, or ... as this is highly coupled to the RecordFactory in use. Why
not have this in the RecordFactory as the current patch, and as part of
YARN-711 a public proxy class is put in place using the current RecordFactory?
Then if you use a different RecordFactory, things will work as expected.
YARN-711, IMO, is not the best place to do this - that deals with how YARN
records are created from their individual arguments. The current implementation
is tightly coupled with the RecordFactory being used, but doesn't necessarily
need to be. I think we need to decide whether this will be a blanket
serialization mechanism for all YARN records, or a selective set.
> Add to ser/deser methods to RecordFactory
> -----------------------------------------
>
> Key: YARN-710
> URL: https://issues.apache.org/jira/browse/YARN-710
> Project: Hadoop YARN
> Issue Type: Bug
> Components: api
> Affects Versions: 2.0.4-alpha
> Reporter: Alejandro Abdelnur
> Assignee: Alejandro Abdelnur
> Attachments: YARN-710.patch
>
>
> I order to do things like AMs failover and checkpointing I need to serialize
> app IDs, app attempt IDs, containers and/or IDs, resource requests, etc.
> Because we are wrapping/hiding the PB implementation from the APIs, we are
> hiding the built in PB ser/deser capabilities.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
