MS Exchange periodically locking up after moving servers to C3750G

toby.mcmichael · ‎03-24-2009

We moved a customers servers running ESX VMWARE from HP Procurve switches to Catalyst 3750G switches at the weekend. After this their MS Exchange Server has locked up 3 times with processes having to be killed and server rebooted to restart exchange. It's been OK since 16:00 yesterday but there is doubt over the new switches ("It was OK until we started using the Cisco Switches"). I can't find any errors on the switches and do not believe there is a network fault, but has anyone seen specific issues with VMWARE / MS Exchange and Cat 3750's. IPbase switches 12.2.46SE.

I'm convinced there is no issue with the switches but the customers is not.

tom_fox_ki4qqq · ‎03-24-2009

Has the customer checked Event Viewer in Windows? If you're not seeing errors on the switches, the next stop would be the servers.

toby.mcmichael · ‎03-25-2009

Thanks Tom.

You are right, the Servers are the most likely suspect, but the customer insisted I look on all forums for any known problems. The Event viewer shows the RPC process going maxed out at around the time of failure and loss of connection to the PDC. Also, other VMWARE services running on the same box are OK, so my feeling is it's not the network but an MS Exchange issue.

Thanks again, Toby.

Kevin Brennan · ‎03-24-2009

Hi Toby,

I havn't seen this specific problem, but I did have issues when moving from a 3com switch to a Cisco switch in a VMWare environment.

What network cards are you using in your VMWare host machine? I'd also check the native VLAN if you are trunking to your VMWare hosts. Some older versions of VMWare ESX did strange things with untagged frames and the native VLAN.

What is your vswitch configuration on the VMWare host? Are any other guest machines affected?

Kevin

toby.mcmichael · ‎03-25-2009

Thanks Kevin.

The Servers are connected to switch ports configured in Access mode. Other Services are not affected on the same VMWARE box.

My strong feeling is this is not a network fault, but I was pushed in to checking the forums just in case.

Thanks again, Toby.