Hi Stuart,
Am 25.02.2017 um 19:56 schrieb Stuart Rench:
I am running haproxy (v 1.7.2 and previously 1.6.x) with nbproc = 4
and only a simple cpu-map in the global section putting 1 process per
core. I am seeing behavior where one or more processes die but are not
regenerated by haproxy master. Is there a known bug?
This is expected behavior; haproxy does not manage processes, there is
no master
process (as opposed to nginx which has a master process).
The general idea is that haproxy processes are not supposed to die. If
they do its
a major bug.
Is the process actually crashing or did you just test it by killing it?
Regards,
Lukas
