Re: [Kea-users] Segmentation fault

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Re: [Kea-users] Segmentation fault

Fohr Jean noel


Same issue here, i'm using the 1.1 deb on debian 8, and have the exact same seg fault with the same log traces. Also tested on Centos7 with EPEL 1.1 rpm with the same result.

I'm connecting to a remote mysql master-master cluster  through a load balancer (active/standby setup)

Didn't get any answer after posting my issue here a few months ago, might be related to the remote mysql part?

Anyway right now i automatically restart the process if it crashes and it seems to be a bypass of the problem, at least under very low load (testing phase)


Jean-Noël Fohr
CHU de Clermont-Ferrand
Equipe Réseaux/Télécoms - [hidden email]

Mail: [hidden email]

De : Kea-users <[hidden email]> de la part de Bill Pye <[hidden email]>
Envoyé : vendredi 23 juin 2017 19:20
À : kea-users
Objet : [Kea-users] Segmentation fault
Hi everyone

I'm having a problem with my kea-dhcp4 server(s) falling over regularly with a segmentation fault, the line I'm seeing in the logs is this:

kernel: kea-dhcp4[28868]: segfault at 451 ip 00007fd8ee50c59c sp 00007ffeb5dada70 error 4 in[7fd8ee4e4000+320000]
systemd: kea-dhcp4.service: main process exited, code=killed, status=11/SEGV
systemd: Unit kea-dhcp4.service entered failed state.
systemd: kea-dhcp4.service failed.

This is the current kea-1.2 built on a CentOS7 (fully updated) server and I also use an updated kernel: 4.11.2-1.el7.elrepo.x86_64 - although I did have this problem on the previous 1.1 release from the EPEL repo. The db is a MariaDB 10.2.6-1 and setup as a Galera cluster. As I mentioned in another recent post, I use Proxysql for access to the cluster but kea doesn't like that configuration so I connect kea directly to one of the db servers.

Has anyone else seen this? If there's any further debug logging you need just let me know.



Kea-users mailing list
[hidden email]