[ 
https://issues.apache.org/jira/browse/STORM-837?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14634566#comment-14634566
 ] 

Arun Mahadevan commented on STORM-837:
--------------------------------------

In HdfsState we support a TimedRotation policy which can make state saving 
complicated. For e.g. if the rotation happens after we record the offset for 
txnid in index file but before any data is written we can lose track of the 
data file.

A few options are :-

1. Support only size based rotation policy to keep it simple.
2. Automatically turn off exactly once in HdfsState (with warnings in logs) if 
TimedRotation policy is chosen.
3. Add another implementation of HdfsState where exactly-once semantics is 
supported and can only take sized based rotation policy.


> HdfsState ignores commits
> -------------------------
>
>                 Key: STORM-837
>                 URL: https://issues.apache.org/jira/browse/STORM-837
>             Project: Apache Storm
>          Issue Type: Bug
>            Reporter: Robert Joseph Evans
>            Assignee: Arun Mahadevan
>            Priority: Critical
>
> HdfsState works with trident which is supposed to provide exactly once 
> processing.  It does this two ways, first by informing the state about 
> commits so it can be sure the data is written out, and second by having a 
> commit id, so that double commits can be handled.
> HdfsState ignores the beginCommit and commit calls, and with that ignores the 
> ids.  This means that if you use HdfsState and your worker crashes you may 
> both lose data and get some data twice.
> At a minimum the flush and file rotation should be tied to the commit in some 
> way.  The commit ID should at a minimum be written out with the data so 
> someone reading the data can have a hope of deduping it themselves.
> Also with the rotationActions it is possible for a file that was partially 
> written is leaked, and never moved to the final location, because it is not 
> rotated.  I personally think the actions are too generic for this case and 
> need to be deprecated.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to