[monit-dev] Resolver issue

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

[monit-dev] Resolver issue

David Kohen
Hello,
We had an issue with our primary internal DNS server, we had to replace it.
I have notices that a heartbeat check is taking about 5 seconds whereas on other servers, it was in the area of 0.01 seconds.
After a restart of the monit daemon, it did the heartbeat in 0.016 seconds, which made me realize that the daemon is caching the nameservers to speed up lookups, but this makes it difficult to do changes in the nameservers without restarting the daemon.
Could this behavior be changed or is it something that could never be changed?

David Kohen
Waze IFS team.


_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev
Reply | Threaded
Open this post in threaded view
|

Re: [monit-dev] Resolver issue

martinp@tildeslash.com
Hello David,

monit doesn't cache the resolved names nor nameservers - it uses the (libc's) getaddrinfo() interface to resolve the hostnames when necessary. Please can you provide more informations about the system so we can try to replicate the problem? (platform, OS, version). Do you use nscd? (name service caching daemon)

Regards,
Martin


On May 1, 2012, at 1:44 PM, David Kohen wrote:

> Hello,
> We had an issue with our primary internal DNS server, we had to replace it.
> I have notices that a heartbeat check is taking about 5 seconds whereas on other servers, it was in the area of 0.01 seconds.
> After a restart of the monit daemon, it did the heartbeat in 0.016 seconds, which made me realize that the daemon is caching the nameservers to speed up lookups, but this makes it difficult to do changes in the nameservers without restarting the daemon.
> Could this behavior be changed or is it something that could never be changed?
>
> David Kohen
> Waze IFS team.
>
> _______________________________________________
> monit-dev mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/monit-dev


_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev
Reply | Threaded
Open this post in threaded view
|

Re: [monit-dev] Resolver issue

David Kohen
In reply to this post by David Kohen
We're using Ubuntu 10.04.4 LTS, x64 (on Amazon AWS).
No NSCD

David Kohen
Waze IFS team.




Date: Wed, 2 May <a href="tel:2012" value="+9722012">2012 11:18:59 +0200
From: Martin Pala <[hidden email]>
To: The monit developer list <[hidden email]>
Subject: Re: [monit-dev] Resolver issue
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=iso-8859-1

Hello David,

monit doesn't cache the resolved names nor nameservers - it uses the (libc's) getaddrinfo() interface to resolve the hostnames when necessary. Please can you provide more informations about the system so we can try to replicate the problem? (platform, OS, version). Do you use nscd? (name service caching daemon)

Regards,
Martin


On May 1, <a href="tel:2012" value="+9722012">2012, at 1:44 PM, David Kohen wrote:

> Hello,
> We had an issue with our primary internal DNS server, we had to replace it.
> I have notices that a heartbeat check is taking about 5 seconds whereas on other servers, it was in the area of 0.01 seconds.
> After a restart of the monit daemon, it did the heartbeat in 0.016 seconds, which made me realize that the daemon is caching the nameservers to speed up lookups, but this makes it difficult to do changes in the nameservers without restarting the daemon.
> Could this behavior be changed or is it something that could never be changed?
>
> David Kohen
> Waze IFS team.
>
> _______________________________________________
> monit-dev mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/monit-dev




------------------------------

_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev

_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev
Reply | Threaded
Open this post in threaded view
|

Re: [monit-dev] Resolver issue

martinp@tildeslash.com
Can you yet provide the content of /etc/nsswitch.conf (just the "hosts:" line) and /etc/resolv.conf (you can obfuscate the real IP addresses of nameservers with "1.2.3.4") from the machine?

We had similar problem report from other user who is however using "mdns" and "nis" nameservice - the problem seems to be the getaddrinfo() implementation and its backend. We were unable to replicate the problem with plain "files, dns" settings.

Regards,
Martin


On May 7, 2012, at 11:27 AM, David Kohen wrote:

We're using Ubuntu 10.04.4 LTS, x64 (on Amazon AWS).
No NSCD

David Kohen
Waze IFS team.




Date: Wed, 2 May <a href="tel:2012" value="+9722012">2012 11:18:59 +0200
From: Martin Pala <[hidden email]>
To: The monit developer list <[hidden email]>
Subject: Re: [monit-dev] Resolver issue
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=iso-8859-1

Hello David,

monit doesn't cache the resolved names nor nameservers - it uses the (libc's) getaddrinfo() interface to resolve the hostnames when necessary. Please can you provide more informations about the system so we can try to replicate the problem? (platform, OS, version). Do you use nscd? (name service caching daemon)

Regards,
Martin


On May 1, <a href="tel:2012" value="+9722012">2012, at 1:44 PM, David Kohen wrote:

> Hello,
> We had an issue with our primary internal DNS server, we had to replace it.
> I have notices that a heartbeat check is taking about 5 seconds whereas on other servers, it was in the area of 0.01 seconds.
> After a restart of the monit daemon, it did the heartbeat in 0.016 seconds, which made me realize that the daemon is caching the nameservers to speed up lookups, but this makes it difficult to do changes in the nameservers without restarting the daemon.
> Could this behavior be changed or is it something that could never be changed?
>
> David Kohen
> Waze IFS team.
>
> _______________________________________________
> monit-dev mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/monit-dev




------------------------------

_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev


_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev
Reply | Threaded
Open this post in threaded view
|

Re: [monit-dev] Resolver issue

David Kohen
The hosts line is "files dns", the resolv.conf file has just the nameservers and their IPs.
We're using version 5.2.5, since there's a patch that I sent that has not been integrated and it's features are crucial for us.

David Kohen
Waze IFS team.



On Wed, May 9, 2012 at 11:08 AM, Martin Pala <[hidden email]> wrote:
Can you yet provide the content of /etc/nsswitch.conf (just the "hosts:" line) and /etc/resolv.conf (you can obfuscate the real IP addresses of nameservers with "1.2.3.4") from the machine?

We had similar problem report from other user who is however using "mdns" and "nis" nameservice - the problem seems to be the getaddrinfo() implementation and its backend. We were unable to replicate the problem with plain "files, dns" settings.

Regards,
Martin


On May 7, <a href="tel:2012" value="+9722012" target="_blank">2012, at 11:27 AM, David Kohen wrote:

We're using Ubuntu 10.04.4 LTS, x64 (on Amazon AWS).
No NSCD

David Kohen
Waze IFS team.




Date: Wed, 2 May <a href="tel:2012" value="+9722012" target="_blank">2012 11:18:59 +0200
From: Martin Pala <[hidden email]>
To: The monit developer list <[hidden email]>
Subject: Re: [monit-dev] Resolver issue
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=iso-8859-1

Hello David,

monit doesn't cache the resolved names nor nameservers - it uses the (libc's) getaddrinfo() interface to resolve the hostnames when necessary. Please can you provide more informations about the system so we can try to replicate the problem? (platform, OS, version). Do you use nscd? (name service caching daemon)

Regards,
Martin


On May 1, <a href="tel:2012" value="+9722012" target="_blank">2012, at 1:44 PM, David Kohen wrote:

> Hello,
> We had an issue with our primary internal DNS server, we had to replace it.
> I have notices that a heartbeat check is taking about 5 seconds whereas on other servers, it was in the area of 0.01 seconds.
> After a restart of the monit daemon, it did the heartbeat in 0.016 seconds, which made me realize that the daemon is caching the nameservers to speed up lookups, but this makes it difficult to do changes in the nameservers without restarting the daemon.
> Could this behavior be changed or is it something that could never be changed?
>
> David Kohen
> Waze IFS team.
>
> _______________________________________________
> monit-dev mailing list
> [hidden email]
> https://lists.nongnu.org/mailman/listinfo/monit-dev




------------------------------

_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev



_______________________________________________
monit-dev mailing list
[hidden email]
https://lists.nongnu.org/mailman/listinfo/monit-dev