Pokazywanie postów oznaczonych etykietą Exchange. Pokaż wszystkie posty
Pokazywanie postów oznaczonych etykietą Exchange. Pokaż wszystkie posty

2025-04-05

change upn for M365 migration

Some company with near 25k identities (above 30k accounts - many persons or identities have few accounts). Now in process of the big jump on clouds - Exchange Online and SharePoint Online.
And Houston, we have a problem.
On premise users using sAMAccountNames (in 3 domains), as You can remember users can have more than one account so only one account should be synchronized to Azure. So we had to use emails as upn (in ADFS) and during synchronization (email is synced as upn to Azure - by AADC/Entra Connect). To complicate the whole picture users can be switched between their accounts and between domains and... last, but not least, account names (sAMAccountNames) could be changed (some old app requirement). So, if emails are uniqe, we could plan to use emails as upn (one of possible scenarios described in Technet/Microsoft). Only one account per user is synchronised (filtering) and it was almost work with Teams, but now we are making the big leap to clouds.
What's wrong? After first steps toward hybrid we have the problem with employees who have access to more than one mailbox. Still, without migrated mailboxes, they start receiving logon request (form based login in Outlook) - expecting to provide valid email address. But, as You can remember - we have separated upn on premise (one of three possible, because we have three domains), but it's different than email addresses and upn on Azure, because upn on Azure is our email, but user is receiveing proposal with on premise upn which is different than email.
To the whole picture You should know, that we had blocked some traffic to outlook.com or outlook.net domains, but we should enable it to proper work of sharepoint online. We had to also set - enforce - in registry to avoid using M365 autodiscover... i think this is the whole picture...
So we can synchronize upn with emails, but:
  • at first look we can recognize few apps with invalid access due to upn change
  • if a user provide valid email outlook will work properly
  • we don't know what will be affected by this change - we have 300-400 apps so impact is unknown

2013-04-16

Windows 2008R2 multi-homed Exchange server loses communication

Problem: Exchange 2010 servers lost communication with domain - with domain controllers
Scenario:
  • multi-homed Exchange 2010 servers - with two network interfaces - 1st to communicate with domain controllers, the 2nd to communicate with users;
  • default gateway is on the 2nd network card - to communicate with users;
  • the 1st NIC hasn't got default gateway - communication with domain controllers is possible via static route;
  • when some of domain controllers are disabled (restart or something) then some of Exchange servers are losing their connection with domain controllers;
  • lost communication means everything available by routes on the 1st NIC is unavailable; static routes are present, but Windows 2008 R2 disallows communication; when I remove these routes and add it again then communication returns;
  • important: firewall is enabled but whole communication is allowed.
Error is repeatable, customer is a big bank with many localizations, during this strange behaviour other systems are working fine - only some of these Exchange servers are losing communication, e.g. 5 of 10 Exchange servers can't communicate with domain controllers and rest of them still can communicate.

Error disappears when firewall is disabled, but company has got policy with requirement of enabled firewall.  

Solution: 
suspected service is Network Location Awareness (NLA) which comes to play when something in networks is changing. It discovers that domain is unavailable (missing domain controller used by Exchange services) and tries to switch network from Domain to... Public but probably it is not possible (error?). 
Similar problem is described here KB980873

We used information available in: Technet:
- in gpo for these servers
- in section: Computer Configuration | Windows Settings | Security Settings | Network List Manager Policies
- change  Location Type from Not configured to Public or Private, but not leave it in default Not configured.

After change from Not configured to Public problem disappears.