Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
None
-
None
-
MXS-SPRINT-146, MXS-SPRINT-147
Description
Recently there have been bugs where the SystemD watchdog kills the MaxScale process due to one of the thread hanging somewhere. This results in a stacktrace that is mostly unhelpful as the signal seems to be delivered to one of the idle threads instead of the thread that actually causes it. To make the stack traces for these cases better, we could use gdb (if installed) to dump all threads.
It should be as simple as executing the equivalent of:
gdb --pid=$(pgrep maxscale) -ex 'thr a a bt' -batch
|