nsd users - Mar 2014 - How to notify myself if NSD dies unexpectedly?

Hello,

I run this big site which has been down for days now without my
knowledge due to this:

[1393533384] nsd[1014]: notice: nsd started (NSD 3.2.15), pid 1014
[1394026594] nsd[1014]: warning: server 1060 died unexpectedly with
status 9, restarting
[1394026594] nsd[1014]: error: fork failed: Cannot allocate memory
[1394113596] nsd[1014]: warning: signal received, shutting down...

Restarted NSD manually just now and everything is fine. But does
anybody know of ways I can notify myself should this happen again?

Thanks!

O.D.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.nlnetlabs.nl/pipermail/nsd-users/attachments/20140306/7a620407/attachment.htm>

On 06/03/2014 14:53, opendaddy at hushmail.com wrote:
> Hello,
> 
> I run this big site which has been down for days now without my
> knowledge due to this:
> 
> [1393533384] nsd[1014]: notice: nsd started (NSD 3.2.15), pid 1014
> [1394026594] nsd[1014]: warning: server 1060 died unexpectedly with
> status 9, restarting
> [1394026594] nsd[1014]: error: fork failed: Cannot allocate memory
> [1394113596] nsd[1014]: warning: signal received, shutting down...
> 
> Restarted NSD manually just now and everything is fine. But does
> anybody know of ways I can notify myself should this happen again?
How about monitoring your server with something like Nagios or Icinga
using their DNS plugins? If you don't want to run a full Nagios or
Icinga server, you might want to look at a service such as Pingdom,
which can monitor things for you and send you reports.

Anand

On Thu, 6 Mar 2014, opendaddy at hushmail.com wrote:
> I run this big site which has been down for days now without my knowledge
due to this:
> 
> [1393533384] nsd[1014]: notice: nsd started (NSD 3.2.15), pid 1014
> [1394026594] nsd[1014]: warning: server 1060 died unexpectedly with status
9, restarting
> [1394026594] nsd[1014]: error: fork failed: Cannot allocate memory
> [1394113596] nsd[1014]: warning: signal received, shutting down...
> 
> Restarted NSD manually just now and everything is fine. But does anybody
know of ways I can notify myself should this happen again?
nsd4 can be started without forking, so an init systems like systemd can
get restarted by them. With systemd that would be Type=Simple

For nsd3, this is not possible due to its forking and mishandling of pid
files.

Paul

On Thu, Mar 06, 2014 at 01:53:18PM +0000,
 opendaddy at hushmail.com <opendaddy at hushmail.com> wrote 
 a message of 54 lines which said:
> But does anybody know of ways I can notify myself should this happen
> again?
IMHO, there is nothing nsd-specific. You just use a generic monitoring
framework. I use Icinga <http://www.icinga.org/> but there are many
others.

Like most tools, Icinga accepts the Nagios API so you can use the
Monitoring plugins (former Nagios plugins)
<https://www.monitoring-plugins.org/> I use an unmodified check_dns,
configured that way in Icinga:

define service{
        use                             generic-service        
        host_name                       mononoke
        service_description             DNS4
        check_command                   check_dig!-4 -l eu.org -T SOA -A +dnssec
        }

define service{
        use                             generic-service        
        host_name                       mononoke
        service_description             DNS6
        check_command                   check_dig!-6 -l eu.org -T SOA -A +dnssec
        }