Il 23/10/2025 08:18, Steven Levine ha scritto:
In <list-14480631@2rosenthals.com>, on 10/19/25
at 12:23 PM, "Massimo S." <ecs-isp@2rosenthals.com> said:
Hi,
Of course no, sometimes it "zombify" even 2 times per day and a reboot is
required. (I've a lot of dumps of the named process in this condition if
this can help). Process unkillable and unresponsive.
Rather than "of course no", I suspect you mean unfortunately no. Dan's
dns servers stay up for weeks at a time. His setup is not huge, but there
are more than 70 zone files.
some minutes ago, i received a notify from my fault-daemon
again bind on MTA1 has become unresponsibe and unkillable
i neither was connected to the vm..
at 9:17 syslog (from bind) wrote this:
...
could not listen on UDP socket: address in use
creating IPv4 inteface failed; interface ignored
not listing on any interface
unable to listen on any interface
loading configuration: failure
exiting (due to fatal error)
this is the fault-daemon that tried to restart bind since it was not
answering dns requests (the fault-daemon do a check each 5 minutes)
so i connected to the VM, i tried to kill it, but the process was unkillable
and i had to setboot /b
what was doing bind before the "unresponsibe+unkillable" state?
nothing special, just answering dns requests
there is no DDoS, i've examined the logs
if this can help, the same bind build on the other VM 2 cores is not showing the issue
(this VM instead has 8 cores assigned), dunno if this is due to SMP
it also do one or two daily scheduled restarts
See that in the past i allways modified zones with bind running and it
never exited/crashed.
The of course make me want to ask, what do you think you changed that
caused this to strat occurring?
Steven
in the latest case i was adding dmark record to a number of domains
but this is not the point, it can happens also for other kind of zone modifies
massimo
|