Hello Vincent,
thanks for your reply. I have done what you said... but there ist nore
core file dumped in /tmp/.
running gdb console it only says "[Inferior 1 (process 719) exited
normally]" when the segfault error occours and haproxy restarts. bt and
bt full results in "No stack."
Can you help about this?
kind regards,
marcus
Am 02.10.2017 um 11:35 schrieb Vincent Bernat:
❦ 2 octobre 2017 10:31 +0200, Marcus Ulbrich <[email protected]> :
I am running haproxy 1.7.9-1~bpo9+1 on debian 9.1. And after running a
while with production data haproxy stops working wiith segmentation
fault:
haproxy[26291]: segfault at 5562af80e000 ip 00007f5985e48149 sp
00007ffe1d613488 error 4 in libc-2.24
Can you please help or have any ideas?
Try to obtain a core file:
- don't use chroot directive in /etc/haproxy/haproxy.cfg
- sysctl -w kernel.core_pattern=/tmp/core.%e.%p.%h.%t
- systemctl edit haproxy.service and put the following:
[Service]
LimitCORE=infinity
- systemctl daemon-reload
- systemctl restart haproxy
Check with "cat /proc/$(pidof haproxy | head -1)/limits" both limits for
"core file size" is unlimited.
Once you get a core file in /tmp/core.haproxy.something, use:
- apt-get install haproxy-dbgsym libc6-dbg gdb
- gdb /usr/sbin/haproxy /tmp/core.haproxy.something
- bt full (post the result)
To reverse your system:
- apt-get remove haproxy-dbgsym libc6-dbg gdb
- rm /etc/systemd/system/haproxy.service.conf.d/something.conf
- systemctl daemon-reload
- sysctl -w kernel.core_pattern=core
