Junping Du commented on YARN-668:

bq. The newly added version id may not be needed. For compatibility issue, we 
can either explicitly check the version id mismatch and throw version mismatch 
exception, or we can just check the specific required field and throw logic 
Agree. Sync up with Vinod offline and he is OK to remove the version id. Will 
remove it soon.

bq.  indentation of the annotation
Nice catch, will fix it soon.

bq. AMRMTokenSecretManager#newInstance() if this method is only used by test, 
we can move it to test?
Make sense. Remove this method as it only use once.

bq. previous comment from Vinod: “The proto definitions need to be in 
server-common. ”
Things will become more complexity if moving proto to server-common for this 
patch. We need to move the proto obj inside TokenIdentifier now to server side 
also (or make hadoop-yarn-common to depend on hadoop-yarn-server which is not 
correct), which means we need to get rid of all getter methods now that affect 
too many places in code base. I would suggest to have a separated refactor 
patch to move proto to server side. Thoughts?

bq. To convert stream into byte array. byte[] buffer = IOUtils.toByteArray(dis);
Make sense. Will replace it.

> TokenIdentifier serialization should consider Unknown fields
> ------------------------------------------------------------
>                 Key: YARN-668
>                 URL: https://issues.apache.org/jira/browse/YARN-668
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Siddharth Seth
>            Assignee: Junping Du
>            Priority: Blocker
>         Attachments: YARN-668-demo.patch, YARN-668-v2.patch, 
> YARN-668-v3.patch, YARN-668-v4.patch, YARN-668.patch
> This would allow changing of the TokenIdentifier between versions. The 
> current serialization is Writable. A simple way to achieve this would be to 
> have a Proto object as the payload for TokenIdentifiers, instead of 
> individual fields.
> TokenIdentifier continues to implement Writable to work with the RPC layer - 
> but the payload itself is serialized using PB.

This message was sent by Atlassian JIRA

Reply via email to