[quagga-dev 10232] Quagga SEGV (possibly due to logrotate SIGUSR1)

Jeroen Massar jeroen at massar.ch
Fri Jan 25 08:07:05 GMT 2013

Package: quagga
Severity: high

High severity as quagga gets killed, which it should not be doing...

Stack trace below.

I guess the below (and thus quagga dying) comes from the logrotate
script, which sends a kill -SIGUSR1 every once in a while.

/etc/logrotate.d/quagga has:
 kill -USR1 `cat /var/run/quagga/$i.pid`

USR1 is 10 though, SIGSEGV is the 11 listed below

It does not happen all the time, but it does happen often enough to be
quite annoying.

Hence having a /etc/cron.d/quagga_restarter with
# /etc/cron.d/quagga_restarter: restart Quagga if is dead

*/1 * * * * root  [ ! -e /var/run/quagga/bgpd.pid ] || [ ! -f /proc/`cat
/var/run/quagga/bgpd.pid`/exe ] && /etc/init.d/quagga restart

to restart bgpd when it dies...



BGP: Received signal 11 at 1359074051 (si_addr 0x0); aborting...
Backtrace for 13 stack frames:

More information about the Quagga-dev mailing list