pamsauto
asked on
Domain Authentication Slow when IP range / subnet different then DC
I am having issues with what I think is Domain Authentication slowing down some client software. The issue on happens when using Intergrated Authentication with SQL server. There is a lot happening so I will explain what happened to get to this point.
Every thing was working fine prior to the next things.
We had three DC's in our domain. One is a new machine Server 2008 r2 64bit, the other two have been in service for 4 years, Server 2003, 32 bit.
We de-promoted on of the older DC last night. It also was a DHCP and DNS Server. Those functions had been migrated to the new server about a week ago with no issue.
Our network has three ip ranges 192.168.2.x 192.168.3.x and 10.192.21.x.
All DC are on 192.168.2.x subnet. All client PC on this subnet work fine.
If I take a client PC that is in the 192.168.3.x subnet and try to access SQL (in 192.168.2.x subnet) every operation takes about 10-30 seconds per DB call.
If I switch from Intergrated Authenticaion to a saved password the app runs at normal speed.
If I access an internal website that is in the 2.x subnet with the PC while it is on the 3.x subnet the website is fast as it usually is.
If I move the client NIC cable to a port on the same switch (we have Cisco with a three VLans) and it is in the 2.x subnet, everything runs at normal speed.
I don't see any issues in the event viewer of the DC's, the client PC, or the SQL Server Box.
Ping times and bandwidth checks all look normal.
The firewalls are turned off on the DC's, SQL box, and the client.
Does anyone have an idea of what I should be looking for?
Every thing was working fine prior to the next things.
We had three DC's in our domain. One is a new machine Server 2008 r2 64bit, the other two have been in service for 4 years, Server 2003, 32 bit.
We de-promoted on of the older DC last night. It also was a DHCP and DNS Server. Those functions had been migrated to the new server about a week ago with no issue.
Our network has three ip ranges 192.168.2.x 192.168.3.x and 10.192.21.x.
All DC are on 192.168.2.x subnet. All client PC on this subnet work fine.
If I take a client PC that is in the 192.168.3.x subnet and try to access SQL (in 192.168.2.x subnet) every operation takes about 10-30 seconds per DB call.
If I switch from Intergrated Authenticaion to a saved password the app runs at normal speed.
If I access an internal website that is in the 2.x subnet with the PC while it is on the 3.x subnet the website is fast as it usually is.
If I move the client NIC cable to a port on the same switch (we have Cisco with a three VLans) and it is in the 2.x subnet, everything runs at normal speed.
I don't see any issues in the event viewer of the DC's, the client PC, or the SQL Server Box.
Ping times and bandwidth checks all look normal.
The firewalls are turned off on the DC's, SQL box, and the client.
Does anyone have an idea of what I should be looking for?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
All great ideas. Here is what has been done.
DHCP was updated to reflect the new DNS server IP address. The DNS that was on the old server was shut off one week prior to demoting the DC.
All of the affected PC's and servers have been rebooted a few times, and all DNS and IP addresses in them are correct.
The routing on out main switch has the IP Helper Address repointed at the new DHCP server also. That change has been in place for a week also before the demoting of the server.
I had stated yesterday that the event logs on the servers and clients we clean. That is not true.
The SQL box is working against the older DC and the clients are working against the New DC. On the old DC I am seeing a lot of these errors.
Pre-authentication failed:
User Name: ASUS-A52F-1$
User ID: STCLOUD\ASUS-A52F-1$
Service Name: krbtgt/STCLOUD.PAMSAUTO.CO
Pre-Authentication Type: 0x0
Failure Code: 0x19
Client Address: 192.168.3.68
Followed by this error.
Service Ticket Request:
User Name:
User Domain:
Service Name:
Service ID: -
Ticket Options: 0x2
Ticket Encryption Type: -
Client Address: 192.168.3.68
Failure Code: 0x20
Logon GUID: -
Transited Services: -
DHCP was updated to reflect the new DNS server IP address. The DNS that was on the old server was shut off one week prior to demoting the DC.
All of the affected PC's and servers have been rebooted a few times, and all DNS and IP addresses in them are correct.
The routing on out main switch has the IP Helper Address repointed at the new DHCP server also. That change has been in place for a week also before the demoting of the server.
I had stated yesterday that the event logs on the servers and clients we clean. That is not true.
The SQL box is working against the older DC and the clients are working against the New DC. On the old DC I am seeing a lot of these errors.
Pre-authentication failed:
User Name: ASUS-A52F-1$
User ID: STCLOUD\ASUS-A52F-1$
Service Name: krbtgt/STCLOUD.PAMSAUTO.CO
Pre-Authentication Type: 0x0
Failure Code: 0x19
Client Address: 192.168.3.68
Followed by this error.
Service Ticket Request:
User Name:
User Domain:
Service Name:
Service ID: -
Ticket Options: 0x2
Ticket Encryption Type: -
Client Address: 192.168.3.68
Failure Code: 0x20
Logon GUID: -
Transited Services: -
I assume there is a configuration change needed within SQL to start using the new DC, but I have no idea what it would be. I know little about SQL; I'm lucky I know how to spell SQL.
ASKER
I found the issue. The DNS Settings were wrong on the old DC and it was pointing at the non existent DNS server. I had assumed that the DC was pulling its config from a reservation in the DHCP server, but it was not, it was configured on the DC itself....
Grrr.. Thanks for the help. I am splitting the points because you both pointed me in the same direction.
Grrr.. Thanks for the help. I am splitting the points because you both pointed me in the same direction.
If DNS resolution fails, then the system will go to a broadcast mode for resolution of the name to IP address. Crossing subnets is an issue for broadcast resolution, causing a delay. I think your issue has to do with name resolution; something hasn't been updated with the new address - it could be server, client, or network component related. Managed Switches and Routers have "helper"s that all the passing of certain traffic that does not normally cross over subnets.