Hello Thomas, 

On Wed, Feb 18, 2026 at 06:28:41PM +0100, Thomas Bätzler wrote:
> I'm using "ip netns exec" in ExecStart to get the program to run in my 
> namespaces.

I tried a similar setup, debian 13/systemd 257.9/haproxy 3.0.11 and haproxy 
didn't crash for me.
# ip netns pids test
1211
1213
# cat /proc/1211/comm
haproxy

> Feb 09 21:37:09 test kernel: haproxy[5433]: segfault at 28f8 ip
> 00007f6763573d33 sp 00007f6762bf25e0 error 4 in
> libc.so.6[96d33,7f6763505000+164000] likely on CPU 9 (core 12, socket 0)
> Feb 09 21:37:09 test kernel: Code: 1f 84 00 00 00 00 00 48 89 f2 31 f6 e9 36
> f6 ff ff 66 0f 1f 44 00 00 41 56 41 55 41 54 55 53 48 89 fb 64 8b 2c 25 d0
> 02 00 00 <8b> 47 10 25 7f 01 00 00 83 f8 43 0f 87 ac 00 00 00 48 8d 1>

This low mem address 0x28f8 could be a NULL ptr dereference, a read of an 
unmapped page.

> I'd appreciate any pointers!

Can you enable core dumps saving and get a backtrace for the segfaulting thread?

Reply via email to