[
https://issues.apache.org/jira/browse/YARN-4665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15130520#comment-15130520
]
Jason Lowe commented on YARN-4665:
----------------------------------
Wouldn't a REST interface follow the same principal? I haven't looked at the
REST API lately, but I'd expect the submission logic to be a POST followed by
GET polling until the state is ACCEPTED or later. If the GET results in a
no-such-app error then the client retries the POST and continues polling. Yes,
this is not the most ideal REST interface design, but unless I'm missing
something it should be functionally equivalent to the RPC path. In either case
the client is going to have to do some kind of retry to handle failovers. Even
with a synchronous interface we can end up with submissions that appear to fail
from the client's perspective but actually succeed (because it was successfully
recorded in the state store before failing to deliver a response to the
client), so it's not just fire-and-forget from the client's perspective.
> Asynch submit can lose application submissions
> ----------------------------------------------
>
> Key: YARN-4665
> URL: https://issues.apache.org/jira/browse/YARN-4665
> Project: Hadoop YARN
> Issue Type: Bug
> Affects Versions: 2.1.0-beta
> Reporter: Daniel Templeton
> Assignee: Daniel Templeton
>
> The change introduced in YARN-514 opens up a hole into which applications can
> fall and be lost. Prior to YARN-514, the {{submitApplication()}} call did
> not complete until the application state was persisted to the state store.
> After YARN-514, the {{submitApplication()}} call is asynchronous, with the
> application state being saved later.
> If the state store is slow or unresponsive, it may be that an application's
> state may not be persisted for quite a while. During that time, if the RM
> fails (over), all applications that have not yet been persisted to the state
> store will be lost. If the active RM loses ZK connectivity, a significant
> number of job submissions can pile up before the ZK connection times out,
> resulting in a large pile of client failures when it finally does.
> This issue is inherent in the design of YARN-514. I see three solutions:
> 1. Add a WAL to the state store. HBase does it, so we know how to do it. It
> seems like a heavy solution to the original problem, however. It's certainly
> not a trivial change.
> 2. Revert YARN-514 and update the RPC layer to allow a connection to be
> parked if it's doing something that may take a while. This is a generally
> useful feature but could be a deep rabbit hole.
> 3. Revert YARN-514 and add back-pressure to the job submission. For example,
> we set a maximum number of threads that can simultaneously be assigned to
> handle job submissions. When that threshold is reached, new job submissions
> get a try-again-later response. This is also a generally useful feature and
> should be a fairly constrained set of changes.
> I think the third option is the most approachable. It's the smallest change,
> and it adds useful behavior beyond solving the original issue.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
