After upgrade to 8.8.8 DNS-cache does not work when zimbra down

[Ace Suares]

I discovered the following:
Since the upgrade from 8.7 to 8.8 the dns resolver doesn't work if zimbra is down.
At least, that is my assessment at the moment.

The server uses 127.0.0.1 as dns resolver. This is zimbra dns cache, right?
When making a backup, I bring zimbra down, do the backup, and bring zimbra up.
Since the upgrade, the server can not resolve the hostname for the backup server!

So:
1. zimbra down
2. backup host can not be resolved, i.e. backup takes 1 second
3. zimbra up
4. some pruning on the backup host ar executed without problem.

This leads me to believe the dns is down during the backup while zimbra is down.

Anyone with the same problem, or can confirm it?

[phoenix]

That sounds about right and it's the reason I use HA DNS servers (PowerDNS Server & caching Resolver) on my LAN and disable the ZCS dnscache.

[L. Mark Stone]

I much prefer the Linux package dnsmasq over Zimbra's dnscache package.

https://www.missioncriticalemail.com/20 ... ion-guide/

Hope that helps,
Mark

[Ace Suares]

Thanks for the answers.
But this is new behaviour in 8.8.8, it wasn't like that in earlier versions. Should I file a bug?

[phoenix]

Thanks for the answers.
But this is new behaviour in 8.8.8, it wasn't like that in earlier versions. Should I file a bug?

Feel free but I wouldn't hold my breath about it getting any traction.

[Ace Suares]

Does anyone know what process runs zimbra-dnscache?
Maybe this one: /bin/bash /opt/zimbra/bin/zmdnscachectl status norewrite

After I bring Zimbra down for the backup, I do a 'pkill -9 zimbra' just 'to be sure'.
Maybe the process now has zimbra in it's name, and thath's why it's killed.

It would make much sense to run a dnscache that didn't work when zimbra is down, right?

[Ace Suares]

Ok, it's clear. dnscache is started by zmcontrol and thus also stopped by it.

Oh my.

[Ace Suares]

zimbra@nayree:~$ ping www.suares.com
PING www.suares.com (136.243.227.216) 56(84) bytes of data.
64 bytes from qwikzite.kukiko.com (136.243.227.216): icmp_seq=1 ttl=59 time=0.612 ms
64 bytes from qwikzite.kukiko.com (136.243.227.216): icmp_seq=2 ttl=59 time=0.978 ms
^C
--- www.suares.com ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.612/0.795/0.978/0.183 ms
zimbra@nayree:~$ zmdnscachectl stop
Stopping dnscache... done.
zimbra@nayree:~$ ping www.suares.com
ping: unknown host www.suares.com

OMG

[Ace Suares]

But there is a solution!

root # su - zimbra -c 'zmdnscachectl start'
Started dnscache: pid 5794
roo # ping www.suares.com
PING www.suares.com (136.243.227.216) 56(84) bytes of data.
64 bytes from qwikzite.kukiko.com (136.243.227.216): icmp_seq=1 ttl=59 time=0.637 ms
64 bytes from qwikzite.kukiko.com (136.243.227.216): icmp_seq=2 ttl=59 time=0.684 ms
^C
--- www.suares.com ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.637/0.660/0.684/0.034 ms

[Ace Suares]

Since the dnscache don't work after stopping Zimbra, even the ./install.sh or upgrade doesn't work. Qanah, what is your take on this?