Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I was wondering if anyone had seen and fixed this one:

We have a number of Windows 2003 SP2 based DCs which normally work well. However, sometimes after a warm reboot they fail to correctly initialise their network connections - the cards are enabled and appear to be working correctly but traffic does not flow over the cards. This is the case on both their LAN and iSCSI interfaces.

A subsequent warm reboot fixes the issue (e.g. they start up normally and work fine). Cold Reboots do not appear to cause the problem. We've seen this issue with different hardware, both in terms of the server platform (different models of Dell PowerEdge) and different network cards (broadcom & Intel based NICs). And of course we've tried different versions of drivers for all the network cards we have tried.

There are no windows or 3rd party firewalls running on the servers.

share|improve this question

1 Answer 1

up vote 1 down vote accepted

Just a shot in the dark here, but what do your switches say is going on? Do you see any entries about the port being shutdown because of attempted mac spoofing or other security issues? Are the MAC addresses in the MAC table after the first (non-working) reboot?

This smells more like a switching problem then a server/host problem.

share|improve this answer
    
Hi Scott, thanks for replying. You could be right, and I'll try and collect data on that basis when I can, but it doesn't smell like it to me. The servers are connected to an iSCSI switch and a LAN which are different physical networks with no connection or route between them. The NICs are different h/w for each connection. The iSCSI n/w is based on Dell switches and the LAN on HP switches. I'll check the data because your suggestion is a very reasonable one but it just doesn't 'smell' like a switch issue to me if its affecting both the LAN and iSCSI connections. –  RobM Mar 11 '10 at 18:06
    
Are you running any kind of virtualization? If so, the hypervisor could be switching your mac address on reboot and that will cause a 20 minute ARP timeout issue for most hosts... –  Scott Lundberg Mar 11 '10 at 21:10
    
We do use virtualisation, but not on the servers in question - they are all bare metal. –  RobM Mar 12 '10 at 13:16
    
Not sure this really solves my problem but its a perfectly reasonable step in general and I don't see a better answer on the horizon. Thanks for your help Scott. –  RobM Apr 8 '10 at 12:49

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.