You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@trafficserver.apache.org by Shaun mcginnity <sh...@gmail.com> on 2013/09/27 16:28:02 UTC
TS-1650 - Inactivity cop should use TRY_LOCK instead of LOCK to avoid deadlock
Hi Brian,
https://issues.apache.org/jira/browse/TS-1650
I'm seeing this deadlock occurring in a production environment.
I see the patch changes to a try lock which makes sense but what happens to
the actual connection that was causing the deadlock, i.e. will it get
destroyed eventually?
It is also unclear to me so far how we get into this state. You say " a
connection may have died while holding on to a mutex this can result in
deadlock. " - could you give an outline how this can happen?
Regards,
Shaun