[
https://issues.apache.org/jira/browse/TS-2053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13715436#comment-13715436
]
Alan M. Carroll commented on TS-2053:
-------------------------------------
I checked the patch and the code it applies to was massively rewritten for
3.3.5 and this patch can't be applied there so I moved it to 3.2.6. Currently,
in HttpTunnel::consumer_handler there is an ink_assert that c->alive is true so
the case mentioned here should generate a crash, not a hung connection. The
correct fix would be to find out how this method is invoked when c->alive is
false, which should never happen.
> When Consumer(User agent)->write_vio.nbytes = INT64_MAX &
> Producer(Cache)->alive = false HttpSM will hung
> ----------------------------------------------------------------------------------------------------------
>
> Key: TS-2053
> URL: https://issues.apache.org/jira/browse/TS-2053
> Project: Traffic Server
> Issue Type: Bug
> Components: HTTP
> Affects Versions: 3.2.4
> Reporter: Bin Chen
> Assignee: Bin Chen
> Fix For: 3.2.6
>
> Attachments: TS-2053.patch
>
>
> When Consumer(User agent)->write_vio.nbytes = INT64_MAX &
> Producer(Cache)->alive = false, HttpClientSession->client_vc will be called
> write_disable. So client_vc won't be handle when vc is timeout or closed.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
