NiC problems

[machinetype]

We have been having problems pinging our nt bdc from the internet until we reboot the machine. After reboot we can ping it for about one half day to 2 days. We have replaced the nic in it putting it in a diffrent pci slot. This would not be a big problem except for this is our vpn machine. It is one of 2 nt4 servs behind a 3com router. Icmp traffic is allowed through. You can ping this address from the other bdc(that is behind the router) which does not make sense because you would think then that it would be an external problem (eg. our isp) but it is not because it gets fixed upon reboot. thanks in advance

 

[OzDog]

Just to clarify..

1. Cannot ping server from outside router, but can ping from inside.
2. Reboot server.
3. Can then ping from outside AND inside.
4. After 12 - 48 hours go back to step 1.

Is this correct?

If so, I would suggest that the problem is a routing issue of some sort. If you can ping from inside but not outside, then it must be.

Can you normally ping other devices from outside? How about when the PDC stops responding? Try both of these scenarios. If you can contact the other devices all the time (including when the PDC is not answreing), then the problem is isolated to the PDC, otherwise it is a general network issue.

General network issue - check the routers, switches, Internet connections and any other devices involved. Sorry, this is not my area of expertise.

If you find the issue is the PDC, re-service pack it and see how you go. If that doesn't help, update the network drivers and re-service pack again.

Still no go?? I need more details (ie server type, NIC type, drivers, service packs etc & what you have tried and it's results too).

Best of luck Cheers,
Sam

Please let members know if you found their posts helpful.

 

[machinetype]

Yes, your steps 1-4 are correct

The routing issue you mention is what I thought it was at first, but why would a reboot fix this? And I have looked at the router config and found no problem.

The two bdc's behind this router(the problem one and the o.k. one are connected through a linksys 5 port workgroup hub to the 3com router.

We have put 6a on it when we started experiencing these problems. The nic is a 3com 3c905cx-tx The computer is a IBM Personal Computer 6287-91u. As for the nic drivers we have just put in this new one and was having the same problem before.

thanks for the help as I know this is a very mysterious problem.

 

[OzDog]

Ok, so it's one computer, do you have a different type of NIC you can try? Unfortunately it's one of those "Elimination" problems now.

Did you re-apply the service pack when you changed the driver? With NT4 you MUST re-apply the service pack when you make any changes to hardware (esp Networking).

Do you have the same card / driver installed on any other machines? Are any of the other BDCs/PDCs the same machine with the same NIC, drivers and software installed??

My next suggestion is to put a completely different model NIC (maybe an Intel, or something, preferably not a 3Com). If you still have the same problem, it's not the NIC and you can re-install the 3C905, it should be OK.

After that, you can try blowing away the IP stack and all the network settings (make sure you know the local Admin password etc first!!). Do this by removing all network cards, services and protocols from the control panel and then from the registry, and then reboot and re-install the network card and settings and Service pack again.

I don't know if you are able, but if it's a dedicated BDC, it may be quicker & easier to re-build. This would pretty much rule out the software (except for faulty drivers - rule this out if the other servers have same card/driver/service pack combination).

If it's not a dedicated BDC (i.e. it's running some other software) this may not be possible.

Let me know. Cheers,
Sam

Please let members know if you found their posts helpful.

 

[JMcAllister]

When your server stops responding to requests from outside, are you still able to ping other addresses outside from this server? If you issue IPCONFIG is a valid gateway address there? If you are unable to ping, and you issue TRACERT to the same address does it make it to the default gateway? at what point does it stop responding?

If it is responding to other machines on the same subnet without a problem, it is most likely not a problem with the NIC, the NIC isn't really going to care if it is getting packets from outside the router or inside, it is going to handle them the same. So, if it is successful inside, it should be successful outside. The only thing that changes from inside to outside is that rather than sending the packet directly to the machine, it will send it to the default gateway (router), to be routed to the correct network.

Because the problem occurs only after a day or two, and it is resolved by a reboot. I would tend to believe that it would be related to DHCP, if you are using it (more specifically the lease renewal); or the configuration of RRAS on the server. You should have a better idea if you look at the steps above to see where it is actually breaking...

Regards,

~ jmcallister

 

[Spirit]

Try this: Get the server working for 1/2 to 2 days (I think you reboot and it works?) then from a command prompt run "Route print" this will give you your routing table. Then after it stops working repeat this and compare the two.

It may be that your server is forgetting some of the routing table, a server I used to have used to forget after each reboot not quite the same as yours but similar-ish!

If it is this then just manually add the route and see what happens it may be that you need to write a batch file to add this every few hours.

Good luck

Iain

 

[machinetype]

Sam, Yes the sp did get reapplied when i changed the nic. The other computer behind this router does have the same nic now, it did not before we switched out the nic on the problem machine. and they both have the newest drivers. Also these 2 machines are the exact same computer. As for blowing away the stack i will try that as a last resort. I will try some of the other answers that were posted and do that if everything else does not work (this machine is a pretty important server in our structure) thanks...

JmCallister, Yes we are able to ping outside addresses with this machine when it stops responding, a gateway is there, we have a static address on this machine (it is a proxy server) I wish i did know where it was breaking, I guess our isp has blocked tracert traffic in because it stops at a router with our isp's name on it, so i can not see where it is stopping on the way in (doing a tracert stops at the same address on the way in on the good server even though the good one is responding to pings) I thought both tracert and ping both used the same protocol so i do not know how they are blocking one but not the other? Thanks...

Iain, This sounds like a good idea, we just rebooted so i will use route print now and save it then compare. I will keep everybody posted for new developments. Thanks...