https://bz.apache.org/bugzilla/show_bug.cgi?id=64848

--- Comment #18 from Matt M <inconceiva...@gmail.com> ---
After a recent upgrade from 9.0.63 to 9.0.75 (with the 2 websocket fixes from
9.0.77 patched in) I am seeing some occurrences of the WsSession Object stuck
on the heap with state OUTPUT_CLOSED in WsHttpUpgradeHandler under
Http11Nio2Protocol.waitingProcessors.

We are in the process of identifying the steps to reproduce but these
occurrences have been rare so far.

Looking at the more recent reworking of WsSession isOpen / isClosed and
surrounding logic from BZ66574 I am wondering if it makes sense for
OUTPUT_CLOSED to also be checked in the isClosed call?   Without a reproducible
test case in place it is difficult to determine via static code analysis, but I
wanted to mention that we are seeing this after our recent upgrade and did not
see it in 9.0.63.

-- 
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org

Reply via email to