GitHub user KurtYoung opened a pull request:
https://github.com/apache/flink/pull/2388
[FLINK-4347][cluster management] Implement SlotManager core
Implement some SlotManager's logic mainly in 4 parts:
* handle slot request initiated by JobMaster
* sync slot status trigger by TaskExecutor's heartbeat
* handle slot request failed at TaskExecutor's side
* handle Resouce(Container) failure (either triggered by some timeout
detector or cluster manager master. e.g. yarn master)
I'm not sure whether these logics been implemented too complex, since the
local status are not always true due to the distributed environment. The
principle i used is:
* Always trust TaskExecutor.
* Generate each action based on the incoming message type (actually, the
interface been called) & current information i have. And update local
information based on the incoming message.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/KurtYoung/flink jira-4347
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/2388.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2388
----
commit d9a055983c3ff7a1c3049c9403c970cf21173061
Author: Kurt Young <[email protected]>
Date: 2016-08-18T07:48:30Z
[FLINK-4347][cluster management] Implement SlotManager core
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
