[Pdns-users] PowerDNS Segfault - Debian

[ops-linux-admins at softcom.biz]

Hello,

We are trying to use pdns as a replacement for bind9 - we are using the
bind (2) backend and just over 57,000 records.  We find that pdns will
segfault randomly on us generally after rediscover command.  Any help in
diagnosing and resolving this problem would be greatly appreciated.


Running environment is as follows:

Linux dns01 2.6.8-3-686-smp #1 SMP Thu May 25 00:54:20 UTC 2006 i686
GNU/Linux

Distro:  Debian Sarge - with powerdns from etch packages

root at dns01:/etc/powerdns# ldd /usr/sbin/pdns_server
        linux-gate.so.1 =>  (0xffffe000)
        libdl.so.2 => /lib/tls/libdl.so.2 (0x4001d000)
        libz.so.1 => /usr/lib/libz.so.1 (0x40021000)
        libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x40035000)
        libm.so.6 => /lib/tls/libm.so.6 (0x40116000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x4013c000)
        libpthread.so.0 => /lib/tls/libpthread.so.0 (0x40148000)
        libc.so.6 => /lib/tls/libc.so.6 (0x4015a000)
        /lib/ld-linux.so.2 (0x40000000)


PowerDNS Version:

* Debian pdns 2.9.20-3  (http://packages.debian.org/testing/net/pdns)
* Have also tried debian source package and recompiled 
with --enable-static-binaries
* Have also tried static binary (DEB) from powerdns.com (2.9.20)


----------- Segfault Dump ------------
Jul 17 10:17:53 dns01 pdns[19288]: Got a signal 11, attempting to print
trace:
Jul 17 10:17:53 dns01 pdns[19288]: /usr/sbin/pdns_server-instance
[0x80a495a] Jul 17 10:17:53 dns01 pdns[19288]: [0xffffe420] Jul 17
10:17:53 dns01
pdns[19288]:
/usr/sbin/pdns_server-instance(_ZN12Bind2Backend17getUpdatedMastersEPSt6
vectorI10DomainInfoSaIS1_EE+0x67)
[0x80ce0f7]
Jul 17 10:17:53 dns01
pdns[19288]:
/usr/sbin/pdns_server-instance(_ZN12UeberBackend17getUpdatedMastersEPSt6
vectorI10DomainInfoSaIS1_EE+0x33)
[0x80a9d13]
Jul 17 10:17:53 dns01
pdns[19288]:
/usr/sbin/pdns_server-instance(_ZN17CommunicatorClass17masterUpdateCheck
EP13PacketHandler+0xa7)
[0x80b7717]
Jul 17 10:17:53 dns01
pdns[19288]:
/usr/sbin/pdns_server-instance(_ZN17CommunicatorClass8mainloopEv+0x11d)
[0x80bb90d]
Jul 17 10:17:53 dns01
pdns[19288]:
/usr/sbin/pdns_server-instance(_ZN17CommunicatorClass12launchhelperEPv+0
x11)
[0x80c9661]
Jul 17 10:17:53 dns01 pdns[19288]: /lib/tls/libpthread.so.0 [0x4014dced]
Jul 17 10:17:53 dns01 pdns[19288]: /lib/tls/libc.so.6(__clone+0x5e)
[0x4022bdee] Jul 17 10:17:54 dns01 pdns[19286]: Our pdns instance
(19288) exited after signal 6 Jul 17 10:17:54 dns01 pdns[19286]:
Respawning Jul 17 10:17:55 dns01 pdns[19427]: Guardian is launching an
instance Jul 17 10:17:55 dns01 pdns[19427]: This is a guarded instance
of pdns Jul 17 10:17:55 dns01 pdns[19427]: UDP server bound to
10.5.23.50:53 Jul 17 10:17:55 dns01 pdns[19427]: TCP server bound to
authoritative:53 Jul 17 10:17:55 dns01 pdns[19427]: PowerDNS 2.9.20 (C)
2001-2006 PowerDNS.COM BV (May 11 2006, 19:40:40, gcc 4.0.4 20060422
(prerelease) (Debian 4.0.3-2)) starting up
Jul 17 10:19:25 dns01 pdns[19427]: Fatal error in control listener:
Guardian exited - going down as well

-- 
Ian Samuel, B.Sc.
Sr. Linux Administrator