Networks

Ask something
Question Locked

What could cause drastic network slowdowns during a server migration?

0 helpful
5 replies

by jonathon.johnson

Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

I have a domain controller that I'm working on decommisioning. It was the first domain controller in the domain. It was also at one time a file and application server, but that was before my time.

Anyway, I've already transfered all 4 roles to the new server, but for some reason, if I remove it from the domain, all hell breaks loose and no one can login without it taking forever or just not working at all. Even be bootup on some stations doesn't complete. What am I missing?

Not sure if it matters but: I did notice that for some reason in computer management, there's ALOT of sessions in the Shared folders but no actual files open. The share was previously the one hosting files for the applications, but those are all, or should be coming from the new server. The pointer to the new file share is in an INI file on each computer and I've verified each and every one (all 400 odd computers). The connections for SYSVOL, NETLOGON, etc are the same as the other servers load, but I don't want to demote it until I'm sure things will go smoothly.

I experieced this yesterday when I simply unplugged the network cable to the old server to see how many systems would be affected and low and behold, it nearly ran the network to a stop.

From the best I can tell, the network cards, processors, switches, memory and hard drive activities are not being hampered....but I'm not a networking person either. So the only other thing I can think of is this server is cranky. Looking at the switch logs, this one is sending quite alot of pause frames (several thousand a minute) which seems a bit odd also since the other interfaces are only sending a hundred or so every hour or so.

This all started after migrateing our entire system from one set of standup servers to Dell PE2950 rack servers.

5 replies

by ---tk---

FSMO Roles
http://support.microsoft.com/kb/324801.

Trouble shoot Domain Controllers.
http://support.microsoft.com/kb/247811
I would run a dcdiag /v >dcdiag.txt on both DCs to see whats going on

I would also check out this site which is full of screen shots, maybe you missed something http://blog.networkfoo.org/?p=125

Domain Controller Diagnostics tool dcdiag.exe
http://technet.microsoft.com/en-us/library/cc776854.aspx

How to use the network diagnostic tool NetDiag
http://support.microsoft.com/kb/321708

Not saying you have done anything wrong, god knows we have all been there at one point or another... Just some stuff you might want to look over to help trouble shoot the issues you are having.

by jamie schnabel

was that your only global catalog server? If so you need to make another domain controller be a global catalog. By default first server is always global catalog server. To verify go to AD sites and services browse to that server and choose NTDS settings and right click. There should be a check box on the general tab. I would recommend having at least 2 depending on size of organization in case of failure.

by jonathon.johnson

Alright, the old server was still the only GC. I have 2 domain controllers not counting the one I'm working on retireing. Could the old server, being the only GC at the time, not being connected have caused all the login failures and whatnot?

by jonathon.johnson

When I was typing, i couldn't recall if it was 4 or 5 roles. I did the role transfers 2 or 3 months ago, but just now got the servers migrated over. Now it's time to retire the old server.

Thanks for the links though, I will definately take a look further. I thought of trying dcdiag but haven't had the time to run it just yet.

I did however take the response below's advice and checked the GC settings and I hadn't yet added another GC to the domain. I figure that may be some of the cause of my issue. I made the other 2 DC's the GC and it almost seems that the switch is actually slowing down a bit. Maybe, just maybe, that will fix some of the issue. I'm going to let them replicate and whatnot for a few hours and try shutting down the old server once again and see if anymore havoc breaks out.

This discussion is closed to new replies.