[ 
https://issues.apache.org/jira/browse/SOLR-8933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15240562#comment-15240562
 ] 

Mark Miller commented on SOLR-8933:
-----------------------------------

bq. the problem has been largely mitigated away already.

Then there should be no reason to introduce this test code path into the 
production code.

I think there is good reason though - to make sure users of this API use it 
properly. If we are going to try and fix this at all, we want to maintain it 
rather than reverting to counting on Jetties internal impl. That is what the 
test mode code is all about. So I think Error is more appropriate than INFO, 
but actually...

Which makes me think, actually, more than logging at Error level, we should 
probably fail an assert there and prevent tests from passing. We are looking to 
enforce this if we can - a logging indicator only would be worst case I think.

We want the build to fail when someone adds some new code and it does not use a 
shield but should.

> SolrDispatchFilter::consumeInput logs "Stream Closed" IOException
> -----------------------------------------------------------------
>
>                 Key: SOLR-8933
>                 URL: https://issues.apache.org/jira/browse/SOLR-8933
>             Project: Solr
>          Issue Type: Bug
>    Affects Versions: 4.10.3
>            Reporter: Mike Drob
>            Assignee: Mark Miller
>         Attachments: SOLR-8933.patch, SOLR-8933.patch, SOLR-8933.patch, 
> SOLR-8933.patch, SOLR-8933.patch, SOLR-8933.patch, SOLR-8933.patch, 
> SOLR-8933.patch
>
>
> After SOLR-8453 we started seeing some IOExceptions coming out of 
> SolrDispatchFilter with "Stream Closed" messages.
> It looks like we are indeed closing the request stream in several places when 
> we really need to be letting the web container handle their life cycle. I've 
> got a preliminary patch ready and am working on testing it to make sure there 
> are no regressions.
> A very strange piece of this is that I have been entirely unable to reproduce 
> it on a unit test, but have seen it on cluster deployment quite consistently.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to