Thanks for the response.  I neglected to mention that we have tried forcing both 10 and 100tx (both half and full duplex) on both the client side as well as the servers. Portfast is currently enabled, I'm unsure about STP. That I will have to look into.

I currently have somebody researching into possible issues with the cisco switches. Are there any other issues that you could possibly see as problematic?  All of the non-Dell computers (as well as the first batch of gx260 small form factor machines) do not experience this strange behavior.  Motherboard revisions appear to be the same, as well as the ghost image used on the machines.  Even the factory install of XP from Dell does the exact same thing.

Again, any further insight you can provide is greatly appreciated.

Try the following:

1. Within the integrated network card's properties, hard-set the Speed and Duplex to 100Mb Full Duplex or 10Mb Full Duplex(depending on the speed of your Cisco switches). Do not allow the cards and switches to attempt to automatically negotiate.
   
2. If you have switches/hubs that use the Spanning Tree Protocol (STP), enable PortFast (for Cisco switches) or disable STP on the port connected to your system.

Did you have any luck with this problem, as I have exactly the same issues.

 

Thanks for any help

 

Martyn

Unfortunately we don't know what was causing it but we were able to scab together a work around.  There were a few of the PC's that worked great, regardless of how long they were used or what programs were installed on them.

For testing I ghosted one of those PC's to the new Dells and it began working properly. 

I ended up taking that PC, stripping it down to what we wanted for our bare ghost image and then batched the rest of the Dells.

The settings of the PC's having the problems had the exact same settings as the Dell that was working properly. We used the exact same port on the switch going to the exact same servers.  No problems ever occured.

We occasionaly have one PC that begins giving us fits (two in the past ~7 or 8 months.  If you use the system restore point from the previous day everything works great. 

End users have limited rights on the local workstation and are not changing any network settings or installing additional programs.

This can be resolved as it appears to be an issue with Windows.  I have heard that some of these issues may be resolved with XP SP2, although I haven't tried installing it to verify  I haven't gotten to the point of comparing the differences between previous registry save points.  The next time this occurs (hopefully never) I will resort to that.

I'm interested to see how our situations are similar.  What is your experience?

Thanks,
Ryan

Message Edited by rjanuary on 04-21-2004 07:09 PM

Good day all -

Currently my company has roughly 30 GX260s and 300+ GX270s Mid form and Mini form.  We are running into the same problem of sorts. Here is the breakdown

Cisco 2950 switches 100Mb full

GX260/270 Win2k Pro

Lockup is generally on 270s I believe this has to do with the new Intel 865G (vs. 845GV in GX260).

Aso IRQ 18 is assigned to the onboard NIC and IDE controller. 

Next time one of your PCs hangs up instruct the user to disconnect the Network cable... The pc may unfreeze

 

-Charlack

Charlack,

Thanks for the reply.  You are completely correct.  This is the exact same situation that we're running into.  My issue with this approach is that the end user should not have to deal with this particular problem.  We also run into re-authentication issues within our tree.  Files open when the user unplugs the network cable are still locked on the server.  When they try to edit these files when they save their work they run into problems.

The part number for the NIC in the PC's does not sound familiar.  Although I haven't looked at the chipset on the mboard in many months, that doesn't look like the model number on our cards.  I know that it is reported under windows as being an "Intel Pro 1000/MT." 

The knowledge of the same IRQ being assigned to both the NIC and IDE controller could point us in a new direction.  I'll have to do further research and find if our batches of 260's/270's are similar.  Thanks for the insight.  I'll reply when further testing is completed.

 

Sorry - I was a bit vague as to the numbers mentioned above.  I was referring to the M/B chipset.

Cheers... and good luck,

Charlack

 

After doing some research in our batches of these PC's, it appears that our primary IDE controller is running on IRQ 14 and secondary running on 15.  The NIC is still running on IRQ 18. 

Both the ghosted PC's as well as the untouched PC's have the same IRQ assignments.  I may have to look into this a little further though.

I do know that both our 260's as well as our 270's in both form factors all display this erratic behavior.

I had wondered if that's what you were talking about.  I did check the ATA controller and it doesn't even appear to have an IRQ assigned to it.  On the resources tab for the controller info, the only things listed are I/O Range and Memory Range.

Is there another way to find this information?

Sorry - My mistake again.  Check the Ultra ATA adapter... Not IDE.

Charlack