Makes no sense - Outlook / Inter-VLAN issue

Unanswered Question
Mar 15th, 2007

Cut-over our previously flat network a year ago and implemented VLANs. We kept the servers in VLAN 98 and the users in VLAN 90 (DHCP scope all setup for WINS/DNS/, ip helper, routing via core, etc..) and everything has been fine for approx. a year now.

A month ago we implemented 2 new Exchange servers and migrated users mailboxes to them... at which point we started seeing this Microsoft message show up on SOME (not all) Outlook clients: "...retrieving data from exchange server... blah,blah" with progression bar, all of which delaying the user for several seconds... sometimes even pointing to the wrong exchange server!

Anyway, someone here took it upon himself to move him and one of the affected users back to VLAN 98 where the servers reside... and it now appears that the problem has disappeared.. ???

To me, this makes no sense... Can anyone explain how/why this would fix this problem?? And if it actually did, what WAS the problem??

Cheers,

Al

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Paul Williams Thu, 03/15/2007 - 09:29

When you say the issue has "gone away", do you mean thats it gone away for all the users or just the two that moved themselves back to VLAN98.

If its the two users only it would suggest that the clients are using broadcasts to find the mail servers (or a service on the servers) and since a router will not usually propogate broadcast traffic - except when you have an ip helper address to point it in the right direction - the broadcast is failing or getting through on a hit or miss basis.

Try putting IP HELPER entries on to the VLAN 98 definition to point to the servers - this may not solve the problem, but will highlight if the clients are using broadcasts for something.

CanEHdian Thu, 03/15/2007 - 10:49

It went away for the 2 users that were moved to VLAN98 (same VLAN as Exchange servers).

But... as I mentioned earlier, there is an ip helper-address command on the core switch's Interface VLAN90 pointing all users to the DHCP server on VLAN98, which in turn provides all users with a DHCP config (IP addy, def/gtwy, WINS, DNS, etc..) providing each user's Outlook client with the ability to connect to their respective Exchange server (WINS converts Exchange server name to an IP address).

So to recap, all users are able to fire up Outlook and connect to their respective mailboxes which reside on their respective Exchange server on a different VLAN. The problem is that several users are experiencing frequent delays with Outlook (typical message of requesting data from Exchange server). Moving 2 users to VLAN98 APPEARS to have stopped the symptoms...

...but what problem has this resolved?

NOTE: it just so happens that the DHCP server IP identified in my IP Helper-Address command is also the old Exchange server... coincidence??

Cheers,

Al

Jon Marshall Thu, 03/15/2007 - 11:36

Al

I'm not a windows expert but when they migrated to new exchange servers at my place we had intermittent connectivity issues.

The way i solved it was to load up ethereal on my laptop and then monitored the link when i was using outlook. What was happening on my PC was that even tho my mail had been migrated i still had links to the old servers as well as the new so it would intermittently time out while it tried to contact the old server.

I'm not suggesting this is your problem but you could try the same thing to troubleshoot

HTH

Jon

CanEHdian Fri, 03/16/2007 - 05:58

Yep, I hear ya Jon... I've been telling people that it must be an 'application layer' problem, but I'm just baffled on how changing VLANs could stop the symptoms... anyway, yes we will surely need to sniff this one out, thanks.

Al

Jon Marshall Fri, 03/16/2007 - 06:10

Al

If it works on the same vlan but not on a remote one sounds like a broadcast vs unicast issue going on.

A sniffer would help with this.

Good luck

Jon

mahmoodmkl Thu, 03/15/2007 - 11:39

Hi

Can u verify weather intervaln routing is working fine.Check weather portfast is configure on u r access ports.Check speed/duplex issues.check mtu sizes.

Thanks

Mahmood

CanEHdian Fri, 03/16/2007 - 06:01

Mahmoood, I already checked all that... everything appears to be fine. I also believe that if such items were a problem (mismatch, etc..), I'd see a lot more people having this issue... especially if inter-vlan routing wasn't working. ;-)

Thanks,

Al

bbaillie Tue, 03/20/2007 - 11:04

The issue is most likely a DNS/WINS/NETBIOS issue, I will explain. The M$ outlook client usually truncates whatever you type in to be just the hostname (FQDN need not apply). This means if the workstation isn't in the same DNS namespace as the email server then the workstation will not find the email server's hostname to IP address translation in its own DNS space and thus will fail. The next thing a workstation will do is attempt netbios broadcast to locate the hostname to IP address translation and if it is in the same subnet, voila it works if not it then queries WINS. Two items to do that will confirm this behavior are "ipconfig /displaydns" to check for DNS resolution and "nbtstat -c" to confirm NETBIOS/WINS name resolution. If DNS issue then fix the DNS or create and ALIAS, if NETBIOS is your only option then get a WINS server, sorry:( Yes I have done M$ work and certs in the past but its not something I am proud of.

Cheers,

Brian

Mike Iversen Fri, 03/16/2007 - 08:24

Did you upgrade to Outlook 2003 clients when you moved to the new Exchange Servers? We see this behavior with Outlook 2003 on our flat network. This doesn't exactly answer your question why moving back to vlan98 fixed the problem, but I bet if you moved more clients back to vlan98 you would start seeing the problem on that VLAN.

Mike

coolpopsun Sun, 03/18/2007 - 23:16

Hi Sir,

Thanx a lot for your support. I am explaning a problem again as I am new

users in cisco, CCNA certified .

I have three switches two 2950 and one 3550 .

3550 is connected to 2950(1st) as a trunk port at both side and another

2950 (2nd) is also connected to 29501st as a trunk.

3550 is in vtp server domain and having 4 vlan---vlan 2 , vlan 3 , vlan

4 and vlan 5.

Now I want in 2950(1st Switch) that 4 pc in vlan 2 and 4 pc in vlan 3.

And in 2950(2nd Switch) 4 pc in vlan 4 and 4 pc in vlan 5 .

When I did all the configuration I found ,not able to ping management IP

also found problem in intervlan routing in switches.

Not able to ping intervlan .What to do I am so confuse that how to do

intervlan routing in switches.

So please provide me solution : As please use any of the Ip address in

switch as u want and please provide me the fully configuration in both

cisco 3550 server and 2950 client configuration with step by step command

as I am trying this last 3 days but not getting any succsess .

Please do the needful . Please provide complete solution with Ip address

as you want to use . step by step configuration in swtiches _______

I will wait for your kind response.

Regards,

coolpopsun

Actions

This Discussion