[Pdns-users] Pdns 2.9.22 stopped treating NOTIFY or manual retrieves when acting as slave

[Florent Lerat]

Le 4 oct. 2010 à 16:16, Florent Lerat a écrit :

> 
> Le 4 oct. 2010 à 15:51, bert hubert a écrit :
>> 
>> Florent,
>> 
>> In 2.9.22 slaving this many domains might actually take *ages*. We've built
>> 2.9.22.x for the people that slave lots of domains, and it has resolved all
>> these issues in huge deployments.
>> 
>> It can be downloaded from
>> http://svn.powerdns.com/snapshots/pdns-2.9.22.x-2.tar.gz
>> 
>> It runs in setups with 350k slave domains.
>> 
>> Please let me know if this solves your issue!
>> 
>> 	Bert
> 
> Thank you Bert,
> 
> We will try this this asap and I'll let you know
> 
> Florent


Bert,

Here are some news. We successfully compiled 2.9.22.x on our Server2 (which is running CentOS 5.5 / Libc 2.5). We are now processing the notify right away and a manual retrieve works like a charm. We patched the server using this ACL patch http://mailman.powerdns.com/pipermail/pdns-users/2006-March/003115.html (we had to tweak it a little to make it work - I'll try to come up with a proper .diff and post it here later).

That's for the great news.

We also compiled and tested it on our Server3 (running RedHat 9 / Libc 2.3) but when we run pdns, it cycle and crash after a few minutes (probably during the startup sequence, when it pulls the soa from all the domains). The trace is shown below :

Oct  5 13:11:59 Server3 pdns[1729]: Got a signal 11, attempting to print trace: 
Oct  5 13:11:59 Server3 pdns[1729]: /usr/local/sbin/pdns_server-instance [0x80d95e8]
Oct  5 13:11:59 Server3 pdns[1729]: /lib/libpthread.so.0 [0x49051d0d]
Oct  5 13:11:59 Server3 pdns[1729]: /lib/libc.so.6 [0x4919d4f8]
Oct  5 13:11:59 Server3 pdns[1729]: /usr/local/sbin/pdns_server-instance(_ZN17CommunicatorClass12slaveRefreshEP13PacketHandler+0x135) [0x80fb1f1]
Oct  5 13:11:59 Server3 pdns[1729]: /usr/local/sbin/pdns_server-instance(_ZN17CommunicatorClass8mainloopEv+0x118) [0x80efb90]
Oct  5 13:11:59 Server3 pdns[1729]: /usr/local/sbin/pdns_server-instance(_ZN17CommunicatorClass12launchhelperEPv+0xe) [0x80f017a]
Oct  5 13:11:59 Server3 pdns[1729]: /lib/libpthread.so.0 [0x4904bae0]
Oct  5 13:11:59 Server3 pdns[1729]: /lib/libc.so.6(__clone+0x57) [0x49257927]
Oct  5 13:12:00 Server3 pdns[20455]: Our pdns instance (12691) exited after signal 6
Oct  5 13:12:00 Server3 pdns[20455]: Respawning

We are considering the possibility of upgrading the libc in order to try and solve the issue. Unless you have another idea of course.

I will keep you posted.

Florent