[jira] [Commented] (YARN-1366) ApplicationMasterService should Resync with the AM upon allocate call after restart

["Rohith Sharma K S (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/YARN-1366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13809382#comment-13809382 ] 

Rohith Sharma K S commented on YARN-1366:
-----------------------------------------

Hi Bikas,
 I have gone through your pdf file attached (YARN-556) and got understand about over all idea behind this subtask.
I have some doubts , please clariffy 

1. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM
>> I understood like, need to reset lastResponseID to 0 and should not clear ask , release , blacklistAdditions and blacklistRemovals. Is am I correct?

2. During RM restart , RM get new AMRMTokenSecretManager. At this time, there will be difference password. Is this handled from RM side during recovery for individual application? Otherwise impact is , heatbeat to restarted RM get fail with an authentication error "passoword does not match"


> ApplicationMasterService should Resync with the AM upon allocate call after restart
> -----------------------------------------------------------------------------------
>
>                 Key: YARN-1366
>                 URL: https://issues.apache.org/jira/browse/YARN-1366
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: resourcemanager
>            Reporter: Bikas Saha
>
> The ApplicationMasterService currently sends a resync response to which the AM responds by shutting down. The AM behavior is expected to change to calling resyncing with the RM. Resync means resetting the allocate RPC sequence number to 0 and the AM should send its entire outstanding request to the RM. Note that if the AM is making its first allocate call to the RM then things should proceed like normal without needing a resync. The RM will return all containers that have completed since the RM last synced with the AM. Some container completions may be reported more than once.



--
This message was sent by Atlassian JIRA
(v6.1#6144)