[
https://issues.apache.org/jira/browse/TEZ-1273?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14065514#comment-14065514
]
Jeff Zhang commented on TEZ-1273:
---------------------------------
Update the state machine and attach the patch
changes of state machine:
* Separate session and non-session into different states.
Session States: SESSION_IDLE, SESSION_RUNNING
Non-Session States: DAG_RUNNING, DAG_SUCCEED, DAG_KILLED, DAG_FAILED
* Add an intermediate state for terminate: TERMINATING. Separate the terminate
into 2 transitions: START_TERMINATE, FINAL_TERMINATE.
** In the START_TERMINATE, it will whether there's DAG running, send DAG_KILL
first if there's one, otherwise, go to FINAL_TERMINATE transition. In this
stage, we could decide whether could do cleanup.
Cleanup case : kill from client side
No-cleanup case: INTERNAL_ERROR, AM_REBOOT
( Won't do cleanup in the ShutDownhook )
** In the FINAL_TERMINATE, it could do the cleanup if necessary (leave it as
placeholder, there's another ticket tracking for this ) and stop all the
services.
Run the MRRSleep in local cluster successfully in session mode, non-session
mode and with recovering.
[~hitesh] Please help review it, will add Unit Test later.
> Refactor DAGAppMaster to state machine based
> --------------------------------------------
>
> Key: TEZ-1273
> URL: https://issues.apache.org/jira/browse/TEZ-1273
> Project: Apache Tez
> Issue Type: Improvement
> Affects Versions: 0.4.0
> Reporter: Jeff Zhang
> Assignee: Jeff Zhang
> Attachments: Tez-1273.patch, dag_app_master.pdf, dag_app_master2.pdf
>
>
> Almost all our entities (Vertex, Task etc) are state machine based and
> written using a formal state machine. But DAGAppMaster is not written on a
> formal state machine even though it has a state machine based behavior. This
> jira is for refactoring it into state machine based
--
This message was sent by Atlassian JIRA
(v6.2#6252)
