[Pdns-users] Issue with communications hanging, version 2.9.17-13sarge1

[Dave Taylor]

Running on Debian server (2.4.27-1-xeon #1 SMP Tue Nov 30 11:53:29 EST 2004
i686 GNU/Linux)

We continue to have an issue where our 'SLAVE' zones are not being updated.
In the past, we've had to restart the pdns-server and everything works for a
few days before it stops again.  The nameserver seems to be running and
answering questions.  A ps -aux shows 3 pdns_server guardian processes and 9
pdns_server-instance guardian processes running.

Before restarting powerdns today, I checked through the last 7 days of logs
for errors.  The only one that stands out, and it showed up multiple times a
day, is "Error sending reply with sendto (socket=5): Invalid argument".
Although, I'm not sure if that is related as I could not find any info
through a Google search on that particular error.

So I attempted to use pdns_control retrieve zonename to update the zone and
got an error stating that the command failed: socket connection refused.

At that point, I initiated a restart (through init.d), and immediately
received the following error:

Guardian is launching an instance
Oct 18 13:33:35 auth01 pdns[5134]: This is module gmysqlbackend.so reporting
Oct 18 13:33:35 auth01 pdns[5134]: This is a guarded instance of pdns
Oct 18 13:33:35 auth01 pdns[5134]: It is advised to bind to explicit
addresses with the --local-address option
Oct 18 13:33:35 auth01 pdns[5134]: binding to UDP socket: Address already in
use
Oct 18 13:33:35 auth01 pdns[5134]: Fatal error: Unable to bind to UDP socket
Oct 18 13:33:36 auth01 pdns[5110]: Our pdns instance exited with code 1
Oct 18 13:33:36 auth01 pdns[5110]: Respawning

It came over and over.  So I attempted to initiate a stop command and
received the message that pdns was not currently running, although a ps -aux
now showed 6 pdns_server guardian processes.  It seems that at least 3 of
them were dead.  I had to kill off the main pdns_server processes and then
start pdns.  At that point, it came up fine and our SLAVE zones started
refreshing again.

As I stated above, this seems to be happening now ( or we're noticing it)
every 3-5 days.

Any thoughts on this would be appreciated or any fingers pointing me in the
right direction.

Thank you.
Dave