Jump to content

Recommended Posts

Posted

So here is the situation:

 

I run a smallish school network (~75 computers, ~150 users) and I have come across a very strange problem. When the internet connection is lost (due to any number of reasons from crappy service from the ISP to cleaning ladies knocking a cord loose) my domain logons take an extraordinary amount of time (5-10min) to complete. This *only* happens when the internet is off. As long as the internet connection is good, logons are processed in a matter of seconds.

 

The setup:

2 x Windows 2003 R2 x64 Domain Controllers both running DNS

1 x Windows ISA 2006 running DHCP

Clients all run Windows XP SP3 Pro. About half the users have mandatory profiles with redirection and half use regular roaming profiles. Both types are affected by this problem.

 

Until recently I was using only one DC that provided both DNS and DHCP services and a firewall (PfSense) out front. Same problems then.

 

My DNS servers are set to forward to the google DNS servers (8.8.8.8, 8.8.4.4) because after much testing, I get much snappier results then by going with my ISP's DNS servers. That being said, I have the same problem when using my ISP's DNS servers.

 

Ok, I am sure I left out some important details that some nice chap will point out for me, but I am hoping someone with a big brain will be able to tell me what stupid mistake I have made here cause this has been bugging me for almost a year now!

 

Cheers!

Posted

we used to have this problem. We traced it to a bit of 3rd party software we were using for intenet monotoriing NetIntelligence. If that couldnt talk to the internet, it would stuff up the machines..

 

Never come accross such bad software in my life before..

Posted (edited)

> we used to have this problem. We traced it to a bit of 3rd party software we were using for intenet monotoriing NetIntelligence. If that couldnt talk to the internet, it would stuff up the machines.

 

I suppose that could also be our problem, but this problem has been going on for more than a year and just recently (a month ago) we moved to a new facility. During the move, I rebuilt the entire system from scratch and changed most of our software and it is still happening. I am more ready to blame myself for a misconfiguration somewhere than on a 3rd party piece of software, but you never know.

 

> Have you tried with each dc pointing to the other for its primary dns?

 

I have not tried, but this was also happening when I had only one DC. I will try and see what happens.

Edited by BaronVonSchtuk
Posted (edited)

Some potential ways of digging deeper into this one:

 

Much of the advice for debugging GPO issues involved increased telemetry from the startup/logon processes, so that seems like a good place to start:

http://support.microsoft.com/kb/250842

http://support.microsoft.com/kb/221833

 

Use Procmon to look to see what the machines are doing during the delay. Trace with the internet up and without it.

How To Enable System Boot Time Logging using Process Monitor Tool

 

If that doesn't show anything conclusive try watching the traffic between a restarting machine and the internet.

 

You could use

Wireshark · Go deep.

or

Download details: Microsoft Network Monitor 3.4

 

Look to see what happens just before or during the pause. Cross reference with the trace from procmon, and perhaps DNS debug logs from your DCs.

 

Some guesses to the cause:

Webdav on the clients using a proxy the far side of your wan?

An app phoning out, perhaps launching from a startup script that is set to run synchronously?

 

 

Good luck.

P.

Edited by psydii
  • Thanks 1

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...