DNS changes no longer propagating

[simonautomatic]

Hi,

Recently I added a new server to my cPanel DNS Cluster. I now have:

server1 - syncronise (WHM 11.40.1)
server2 - syncronise (WHM 11.40.1)
ns1 - cPanel DNS-only - read only (WHM 11.38.2)
ns2 - cPanel DNS-only - read only (WHM 11.38.2)

Server2 will replace Server1 completely over the next few weeks, as i migrate accounts across. However something odd is happening, as when I update DNS to point MX, www, A records etc to the new server, the DNS is being syncronised across the cluster yet even a day later there is no propagation across the internet, and the sites/email etc are still served from the old server.

Typically I have TTLs of 14400, so I would expect a few hours delay, but not more.

This appears to be a new problem, I'm not sure if it's something I've messed up in the cluster config or what.

I even SSH'ed into the two name servers to check that the zonefiles had updated - which they have. But all attempts to query the domains via dig, or intodns.com show the old DNS.

Confused!

 

[simonautomatic]

It seems that no changes of any kind are propagating with my DNS any more. Very weird.

I've never seen anything like this. Doesn't matter whether I add new A records, change existing ones. Nothing is visible to the outside world.

 

[cPanelMichael]

Hello

You could try restarting the Named service on the DNS-Only machines and rebuilding the DNS configuration via:

/scripts/rebuilddnsconfig

If the issue persists, could you open a support ticket using the link in my signature so we can take a closer look? You can post the ticket number here so we can update this thread with the outcome.

Thank you.

 

[simonautomatic]

Many thanks - ticket created: 4555141

 

[simonautomatic]

Disappointing response so far. A cPanel guy spent hours looking at my servers, and basically came up with nothing (well - my sync settings were wrong, but to me that does not explain the total lack of propagation).

I am seeing some domains propagate very slowly (between 24-48 hours which is crazy when all records are on a TTL of 14400 or less).

Something's badly wrong here and I wish I could fix it.

 

[cPanelMichael]

To update, it looks like both hosting servers contained the same DNS Zones. Thus, they were sending conflicting records to the DNS-Only servers. The solution here is to modify how the cluster is setup, or to disable clustering for one of the hosting servers.

Thank you.