[Pdns-users] Crash with signal 11

[Oli Schacher]

I wrote a script to mass-notify a few domains:

#!/bin/sh
for domain in $* ; do
echo $domain
pdns_control notify-host $domain <slave1-ip>
pdns_control notify-host $domain <slave2-ip>
done ;

We have two slaves running on Centos 5, using sqlite3 backend. When I 
ran this script, using about 40 domains, one slave  crashed:

Apr  7 11:11:59 ns1 pdns[13207]: Created new slave zone '<domain>' from 
supermaster <supermasterip>, queued axfr
Apr  7 11:11:59 ns1 pdns[13207]: Got a signal 11, attempting to print trace:
Apr  7 11:11:59 ns1 pdns[13207]: /usr/sbin/pdns_server-instance [0x80b8cc4]
Apr  7 11:11:59 ns1 pdns[13207]: [0xe63420]
Apr  7 11:11:59 ns1 pdns[13207]: /lib/libc.so.6(memcpy+0x1c) [0x29e6ec]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/lib/libstdc++.so.6(_ZNSs4_Rep8_M_cloneERKSaIcEj+0x64) [0x4a8da4]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/lib/libstdc++.so.6(_ZNSsC1ERKSs+0x57) [0x4a9707]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZNK5QType7getNameEv+0x87) [0x8076447]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN12UeberBackend8cacheHasERKNS_8QuestionER17DNSResourceRecord+0x83) 
[0x80c4d13]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN12UeberBackend6lookupERK5QTypeRKSsP9DNSPacketi+0x340) 
[0x80c5930]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN13PacketHandler11makeCanonicEP9DNSPacketS1_RSs+0xa0) 
[0x8082130]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN13PacketHandler17questionOrRecurseEP9DNSPacketPb+0xa92) 
[0x8087672]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN13PacketHandler8questionEP9DNSPacket+0x29) 
[0x808b069]
Apr  7 11:11:59 ns1 pdns[13207]: 
/usr/sbin/pdns_server-instance(_ZN11DistributorI9DNSPacketS0_13PacketHandlerE10makeThreadEPv+0x3b8) 
[0x80fc418]


I was not able to reproduce this (running the script again after 
restarting the slave).  Is this something to worry about/should I create 
a bug report or did I just make something stupid/hit a known prob?

Thanks,
Oli