I can reproduce this bug with package:
bind9 1:9.8.4.dfsg.P1-6+nmu2+deb7u4
I have several internal networks and a PPPoE client running.
The trigger is when the PPPoE link fails and the interface ppp0 goes
away, along with its local address.
I have discovered a second trigger that is when ppp0 re-appears, bind9
crashes in the same way.
I have further refined this to "if any interface is downed, bind will
crash" - eg:
ifdown eth0.1002
The cause2trigger seems instant and 100% reproducible.
The crash causes bind to syslog the following:
Apr 9 12:45:10 shinybob named[12196]: received control channel command
'reconfig'
Apr 9 12:45:10 shinybob named[12196]: loading configuration from
'/etc/bind/named.conf'
Apr 9 12:45:10 shinybob named[12196]: reading built-in trusted keys
from file '/etc/bind/bind.keys'
Apr 9 12:45:10 shinybob named[12196]: using default UDP/IPv4 port
range: [1024, 65535]
Apr 9 12:45:10 shinybob named[12196]: using default UDP/IPv6 port
range: [1024, 65535]
Apr 9 12:45:10 shinybob named[12196]: no IPv6 interfaces found
Apr 9 12:45:10 shinybob named[12196]: no longer listening on 81.2.78.28#53
Apr 9 12:45:10 shinybob named[12196]: sizing zone task pool based on 46
zones
Apr 9 12:45:10 shinybob named[12196]: view.c:467: REQUIRE(targetp !=
((void *)0) && *targetp == ((void *)0)) failed, back trace
Apr 9 12:45:10 shinybob named[12196]: #0 0xb76ef724 in ??
Apr 9 12:45:10 shinybob named[12196]: #1 0xb72472b4 in ??
Apr 9 12:45:10 shinybob named[12196]: #2 0xb75e66a5 in ??
Apr 9 12:45:10 shinybob named[12196]: #3 0xb75e8700 in ??
Apr 9 12:45:10 shinybob named[12196]: #4 0xb76d2a09 in ??
Apr 9 12:45:10 shinybob named[12196]: #5 0xb7705805 in ??
Apr 9 12:45:10 shinybob named[12196]: #6 0xb77075a9 in ??
Apr 9 12:45:10 shinybob named[12196]: #7 0xb7707dbc in ??
Apr 9 12:45:10 shinybob named[12196]: #8 0xb76e82bc in ??
Apr 9 12:45:10 shinybob named[12196]: #9 0xb76eaf68 in ??
Apr 9 12:45:10 shinybob named[12196]: #10 0xb726a282 in ??
Apr 9 12:45:10 shinybob named[12196]: #11 0xb721ec39 in ??
Apr 9 12:45:10 shinybob named[12196]: #12 0xb7036c6e in ??
Apr 9 12:45:10 shinybob named[12196]: exiting (due to assertion failure)
I think this needs to be raised to a higher priority. I have no elegant
workaround - I have a kudgey workaround of running "monit" to watchdog
bind9 and restart when needed.
I have supply additional config on demand, but please be warned that my
bind9 config is non trivial. It may be worth just trying the ifdown test
on the same version of bind with any running config.
Kind regards,
Tim
--
Tim Watts
Personal Blog: http://squiddy.blog.dionic.net/
http://www.sensorly.com/ Crowd mapping of 2G/3G/4G mobile signal coverage
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
