[AMQ-9683] java.io.IOException on idle HTTP/XA connection#1414
[AMQ-9683] java.io.IOException on idle HTTP/XA connection#1414jbonofre merged 1 commit intoapache:mainfrom
Conversation
|
I think some tests can break with this change. I will double check. |
Yes, I saw that but I would not reproduce them locally. Any ideas? We are currently forking and building our own binary, which is a bit of a pain to keep up with all the dependencies, so we really appreciate your insight here! And thanks in advance! |
|
@jbonofre any luck here? |
|
@colinwerner on it 😄 |
|
I'm checking one test that can be problematic (but not a blocker). |
cshannon
left a comment
cshannon
left a comment
There was a problem hiding this comment.
I think this makes sense, if the transport is already stopping or stopped then exceptions are expected so probably best to log it vs throwing an exception, as long as the empty response is also handled without issue on the return and the tests are all passing.
|
IMO: This is the ideal scenario to add a counter metric. Logging can get clogged up/rotated/etc and having a counter makes it easy to monitor if this is happening in a on-off, or consistently. |
|
I think this first PR is fine as a quick fix. We can always improve later. The debug log message is good enough for now. |
Co-authored-by: colin <[email protected]> (cherry picked from commit 9c3b38a)
|
Thank you @jbonofre -- much appreciated! I'll follow-up if we see any other issues! |
4 participants
As described in https://issues.apache.org/jira/browse/AMQ-9683 an EMPTY payload was received in response to a GET, as an intermediate POST sent a shutdown.
This PR implements solution 1, where we ignore the exception IFF we are shutting down or we are shutdown.
Solution 2 is in #1413. We only need one or the other...