Slave dns server not receiving MX record updates

Hi Guys,

I had a client tell me today that they can't receive emails randomly from their clients. I did some searching and found that the MX records for their domain on my NS1 machine are correctly pointed at google servers and that my slave or ns2 server does not have that mx record information and is still pointing at the a record auto generated during virtualmin setup. When the transfer takes place from the primary to the slave I get the following message in the log on the slave:

Dec 28 14:47:22 venus named[32147]: transfer of 'pleasantsmiledentalcare.com/IN' from 69.33.169.228#53: connected using 69.33.169.230#57269
Dec 28 14:47:22 venus named[32147]: transfer of 'pleasantsmiledentalcare.com/IN' from 69.33.169.228#53: Transfer completed: 0 messages, 1 records, 0 bytes, 0.001 secs (0 bytes/sec)

Attached are the mx records from primary and slave so you can see the difference.

I am baffled and hope that it is a simple fix.

~Jeremy

Status: 
Closed (fixed)

Comments

If you add a record on the master to the zone using Virtualmin, does it get updates on the slave?

If no, what gets logged to /var/log/messages by named on the master and slave systems? You should see some message indicating why the transfer failed ..

Hi Jamie,

/var/log/messages does not generate any output when I add a new a record to the master zone. However the /var/log/syslog does show activity and displays this:

master :

Dec 28 23:01:31 mars named[11749]: client 69.33.169.230#39738: transfer of 'pleasantsmiledentalcare.com/IN': IXFR ended

slave :

Dec 28 23:01:31 venus named[4358]: zone pleasantsmiledentalcare.com/IN: Transfer started.
Dec 28 23:01:31 venus named[4358]: transfer of 'pleasantsmiledentalcare.com/IN' from 69.33.169.228#53: connected using 69.33.169.230#39738
Dec 28 23:01:31 venus named[4358]: transfer of 'pleasantsmiledentalcare.com/IN' from 69.33.169.228#53: Transfer completed: 0 messages, 1 records, 0 bytes, 0.001 secs (0 bytes/sec)

No errors.

What is very interesting is if I delete the slave zone for this domain and then create a new slave zone with the proper master and apply configuration the same (bad) old entries are then pulled from the master and written to the slave zone .hosts file instead of what is currently in the master zone .hosts file. (This is seemingly a cached behavior but I am unable to track it down. I would expect that the current .hosts file for the master zone would simply be duplicated to the slave .hosts file.

Here is the master zone .hosts file:

$ORIGIN .
$TTL 38400      ; 10 hours 40 minutes
pleasantsmiledentalcare.com     IN      SOA     mars.xtsgalaxy.net. root.mars.xtsgalaxy.net. (
                        1320884275
                        10800
                        3600
                        604800
                        38400 )
                        NS      ns2.xtsgalaxy.net.
                        NS      mars.xtsgalaxy.net.
                        A       207.145.224.27
$ORIGIN pleasantsmiledentalcare.com.
_domainkey              TXT     "t=y\; o=-\;"
$ORIGIN _domainkey.pleasantsmiledentalcare.com.
2011                    TXT     "k=rsa\; t=y\; p=MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDXeAQSOWSP0BkihzmrDJn8XibheLPkxA6OQEgeP5uZEApqUX/6M5CsOm7Azbhqsksy37ZKJBg5inK2ObOrqzh+C9qAsx3fJmxpzIQbTDTpMJFU8LLTXhfAtLySuOVTeC+T+jo2uW0uruILdm8OkR84H29DhtZm5vuJsuKwiLTO0wIDAQAB"
$ORIGIN pleasantsmiledentalcare.com.
admin                   A       207.145.224.27
localhost               A       127.0.0.1
www                     A       207.145.224.27
pleasantsmiledentalcare.com.    IN      A       207.145.224.27
pleasantsmiledentalcare.com.    IN      MX      1 ASPMX.L.GOOGLE.COM.
pleasantsmiledentalcare.com.    IN      MX      5 ALT1.ASPMX.L.GOOGLE.COM.
pleasantsmiledentalcare.com.    IN      MX      5 ALT2.ASPMX.L.GOOGLE.COM.
pleasantsmiledentalcare.com.    IN      MX      10 ASPMX2.GOOGLEMAIL.COM.
pleasantsmiledentalcare.com.    IN      MX      10 ASPMX3.GOOGLEMAIL.COM.
ftp.pleasantsmiledentalcare.com.        IN      A       207.145.224.27

Slave .hosts file

$ORIGIN .
$TTL 38400      ; 10 hours 40 minutes
pleasantsmiledentalcare.com IN SOA mars.xtsgalaxy.net. root.mars.xtsgalaxy.net. (
                                1320884270 ; serial
                                10800      ; refresh (3 hours)
                                3600       ; retry (1 hour)
                                604800     ; expire (1 week)
                                38400      ; minimum (10 hours 40 minutes)
                                )
                        NS      ns2.xtsgalaxy.net.
                        NS      mars.xtsgalaxy.net.
                        A       207.145.224.27
                        MX      5 mail.pleasantsmiledentalcare.com.
                        TXT     "v=spf1 a mx a:pleasantsmiledentalcare.com ip4:69.33.169.228 ip4:207.145.224.27 ?all"
$ORIGIN pleasantsmiledentalcare.com.
_domainkey              TXT     "t=y\; o=-\;"
$ORIGIN _domainkey.pleasantsmiledentalcare.com.
2011                    TXT     "k=rsa\; t=y\; p=MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDXeAQSOWSP0BkihzmrDJn8XibheLPkxA6OQEgeP5uZEApqUX/6M5CsOm7Azbhqsksy37ZKJBg5inK2ObOrqzh+C9qAsx3fJmxpzIQbTDTpMJFU8LLTXhfAtLySuOVTeC+T+jo2uW0uruILdm8OkR84H29DhtZm5vuJsuKwiLTO0wIDAQAB"
$ORIGIN pleasantsmiledentalcare.com.
admin                   A       207.145.224.27
ftp                     A       207.145.224.27
localhost               A       127.0.0.1
m                       A       207.145.224.27
mail                    A       207.145.224.27
webmail                 A       207.145.224.27
www                     A       207.145.224.27

~Jeremy

Do you perhaps have more than one slave DNS server? If the wrong records are being transferred, perhaps they are coming from another slave instead of the master..

Hi Jamie,

Well it is official, there is no explanation for this : ).

To get around this issue for tonight I manually copied the contents from the master .hosts file into the slave .hosts file.

I have looked high and low for the older .hosts settings on the slave and master and found nothing. I have no idea where this old content for the slave .hosts file is coming from and why changing the master zone and applying the zone doesn't result in modifications to the slave .hosts file when the log says Transfer completed 1 records.

Is there a way to setup a confirmation in Webmin Bind that the transfer was successful?

I would be happy to let you jump on my servers so you can see what I am talking about.

~Jeremy

Sure, a remote login to your system would be really useful. You can email me details at jcameron@virtualmin.com .. I'd need root SSH access to both the master and slave systems though.

Ok, I fixed the issue - there were two named processes running on your master system, and restarting BIND only killed one of them! The fix was to stop BIND, kill all named processes, then start BIND again.

Thank you Jamie,

Wow, I am shocked I didn't think to check for that!!!

You are the man once again!

Hope you have a Happy New Year

~Jeremy

Automatically closed -- issue fixed for 2 weeks with no activity.