Registered server and new domain name

[JLDT]

Hello, We are using registered servers to allow the connectivity with our differents domains. Currently, we are moving from one datacenter supplier to another and during this operation, we are changing the DC from one supplier to the other. When we stop the old DC, we have issues within ePO as the domain name used in the registered server is still using the old DC. Example : we have one server with the domain name "toto.enterprise.net" (DNS style) This domain name is configured to target some DCs. During the night, we are replacing old DCs with new (so new IP, as we are changing from supplier). Even if we have the right information with NSLOOKUP on the ePO server, it seems the information within ePO are not updated. So all the LDAP queries are failling as the old target is shutdowned. Do you know if there is a task, a file which can be edit to force the refresh within ePO ?

[cdinet]

In the nic properties of the epo server, what servers are listed as the dns servers?  Are any of them the ones removed?  If so, you might need to change that to one that won't be changing, or one that has already changed.  You can also try ipconfig /flushdns and ipconfig /registerdns to see if it picks up the new dc's.

[JLDT]

Hello,

Thanks for your reply.

On the ePO server, DNS have not changed. They are using DNS of the new supplier.

Support team has got the issue yesterday and they have use the ipconfig /flushdns to refresh. I can't confirm they have run the ipconfig /registerdns. 

They were able to reach the new DC with Nslookup.

It's only ePO which was not using the new DNS to contact their active directory (so they were not able to log on the ePO application). It looks like if there was a cache somewhere in ePO. 

Do you think that the User LDAP Sync can solve the issue ? (i mean, if i'm able to run the server task, does it help ?)

Thanks again for your help.

[cdinet]

I would check the registered servers in epo for the ldap server.  Any time there is a break in communication with it, then it can affect that registration.  You can verify that by seeing if any listed in registered servers no longer show up in server settings, active directory groups.  If they are no longer there, then go back to registered servers, validate the test connection is successful and re-save the entry, even if you don't make any changes.

[JLDT]

Thanks again.

That's what we are doing. We are going to the registered server, then change from the DNS style to server name or even IP. Keeping this settings for few hours (half day because we are working on something else) and change back to the DNS style.

Works like a charm with this. 

But why do we need to replace DNS by servername/IP ? ePO application is not using the DNS info of the ePO Server ?

Thanks again for sharing your thought.

[cdinet]

It will use dns of local server for some functions, but will use the registered server for others, such as ldap user sync and active directory sync. 

[JLDT]

So if the registered server is using a DNS style entry to reach AD, editing/saving without any change should fix the issue (kind of refresh in ePO) ?

Thanks for your help.

[cdinet · Accepted solution]

Yes, it should, especially if it has gone from the server settings active directory groups.

[JLDT]

Thanks for the explanation ! 

Regards,