Ilya,

Am 02.02.20 um 18:52 schrieb Илья Шипицин:
> Feb 02 20:50:07 xxx haproxy[2851]: [NOTICE] 032/205007 (2851) : New worker
> #1 (2897) forked
> Feb 02 20:50:07 xxx haproxy[2851]: [ALERT] 032/205007 (2851) : Current
> program 'api' (2896) exited with code 1 (Exit)

The 'api' program (i.e. the dataplane API) died with code 1 ...

> Feb 02 20:50:07 xxx haproxy[2851]: [ALERT] 032/205007 (2851) :
> exit-on-failure: killing every processes with SIGTERM

... leading to HAProxy terminating every worker ...

> Feb 02 20:50:07 xxx haproxy[2851]: [ALERT] 032/205007 (2851) : Current
> worker #1 (2897) exited with code 143 (Terminated)
> Feb 02 20:50:07 xxx haproxy[2851]: [WARNING] 032/205007 (2851) : Former
> worker #1 (2854) exited with code 143 (Terminated)
> Feb 02 20:50:07 xxx haproxy[2851]: [WARNING] 032/205007 (2851) : Former
> program 'api' (2853) exited with code 0 (Exit)
> Feb 02 20:50:07 xxx haproxy[2851]: [WARNING] 032/205007 (2851) : All
> workers exited. Exiting... (1)

... leading to the master stopping itself.

> Feb 02 20:50:07 xxx systemd[1]: haproxy.service: main process exited,
> code=exited, status=1/FAILURE
> Feb 02 20:50:07 xxx systemd[1]: Unit haproxy.service entered failed state.
> Feb 02 20:50:07 xxx systemd[1]: haproxy.service failed.

... leading to the unit being stopped.

So you would have to find out with the dataplane API dies. My educated
guess would be that it fails because the port is already being in use by
the old dataplane API process.

Best regards
Tim Düsterhus

Reply via email to