0

I'm running 4 of Win2k3 64bit servers in the same subnet. It's been more than an year that I've running them without a problem.

Recently, I kept losing the connection to one of the server. Let's say it's 'server A' which has a problem. Losing the connection means that I can't access to server A from the other servers. I've checked if server A has any internet connection problems or are there any abnomal event logs in the eventvwr - but haven't found any problems.

The problem usually resolved if I restart the server again. But as time goes by, it keeps happen again and again. I can't afford to restart the server every time, and I really want to find out the reason.

Can anyone help me out? Let me know if you guys need any of more information.

2
  • Single NIC on Server A? Jun 1 '09 at 13:18
  • Every server has double NIC. For public, and private.
    – Sam
    Jun 1 '09 at 13:44
1

If you've got a managed switch, look at the switch logs, port statistics and port link state to identify/confirm the symptoms during the failure.

Faults with autonegotiation, bad cabling, mismatched speed/duplex and framing errors are easy to pick up with this info.

"show log" and "show interface <interfacename>" will do the trick on Cisco gear, I presume managed switches from other vendors will have similar commands.

0
1

I agree with both Marko and Nick, I've had Servers that were "going bad" to be just the switch. As soon as I tested a new switch it was working all fine. In my opinion either the NIC card is going or the switch is going.

1
  • I've had the same experience as Hondalex. Sometimes it's the switch, sometimes it is a NIC. Sometimes I've tracked back to find that a driver update screwed things.
    – NotMe
    Apr 10 '10 at 16:27
0

You could have a name resolution problem or a duplicate name/ip address.

1
  • I've been checked them already but haven't found any duplicate name/IP in the subnet.. Thanks anyways.
    – Sam
    Jun 1 '09 at 13:10
0

Ensure no viruses/rookits are on the server 1st off. After that, see if the NIC is dying. You can run a continuous ping on the server outbound, script it to dump to a text file (ping -t TargetIP >> C:\ping.txt), and let it run for a bit. See if the ping traffic dies suddenly then comes back. Could be indicative of a bad NIC.

0
0

Even though the problem seesm to resolve itself with a restart, don't rule out the possibility of an issue with your switch - maybe try a different port?

Edit: Sorry - duplicate of Nick's post

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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