I searched the forum and did not find any matches to my problem. My EX475 has been experiencing system errors (BSOD):



This MS article points to a NIC problem (0x7c = BUGCODE_NDIS_DRIVER). I upgraded the SiS drivers to version 2.0.1039.1110 but this did not help, the crashes kept occurring. After reading comments on my blog post (obviously I am not the only one) I swapped my DI-624 router (this thing is notorious for rebooting) for something else. The last system error occurred right at the time the router rebooted and I believe that the router behavior could be the cause of the server crashes. Since the swap Feb 28 I have not yet noticed a system error, but I don't feel that the problem has been solved. A router should not be able to take down a server.

This is a boring, plain-vanilla EX475 with no add-ins and other software/services. Just Power Pack 1 and the HPMSS-1.3-R1.EXE update (crashes occurred before as well).

My question is: has anyone else experienced system crashes similar to this one? This really seems to be related to the network card and router issues.

Has anyone tried installing an external USB-type NIC? This would allow me to disable the on-board SiS card. How reliable are these, I wonder.

Thanks.

Did you upgrade the system memory in the 475 recently?

No, memory is still the standard 512MB.

Other than the system error in the event log, I did not see anything else that seems related, i.e. no errors right before the crash. At the time of the last crash that I witnessed (other than that they always seemed to occur when I wasn't around) the server was not overly busy either.

Interesting, I went looking and see that my EX475 had this same issue today. I'll look at the logs more to see if it's happened other times but I do not believe it's been a recurring problem for me.

_________________
Alex Kuretz
Remote Notification: Monitoring and alerting for your Home Server via email, Twitter, text message, RSS, push notifications, and on your Mac desktop
Have you tried our Add-Ins?

I just saw that my Trendnet router (the one I replaced the D-Link with) rebooted this morning for the second time since the swap. It did not affect the server. I should run with this configuration for a few more days and then swap the routers back to see if the problem re-occurs.

The question might be: why is your router rebooting? It's not normal behavior for those devices. I know the only time my router reboots is when i upgrade the Firmware..

Poor Power? Cabling? could be a number of things..

Don't get me wrong but the reason the router reboots is irrelevant, it should not take the server down.

As I stated, the DI-624 is notorious for this. My current router is not powered through a UPS so minor power fluctuations can easily cause this.

Not taking you wrong at all! It is quite possible for a malformed packet to lockup a network card and crash a kernel (esp with a poorly written driver *cough cough SIS*) Home routers are not enterprise style components and are subject to all sorts of quirks. The only reason I was exploring the stability of the router was to potentially reveal the source of the packet/traffic that is causing the server to die. The router crashing all the time could indicate that it's in the same boat as the server, and receiving erroneous packets and not knowing what to do with them. Multiple DHCP servers, might do it. A attack from outside the network could do it too. You're going to have to sleuth a little bit and watch the traffic in your local network and watch what is going on. http://www.ethereal.com now known as wireshark could help in this. You'll be surprised how much traffic goes through your network.

Hi,

I have the same problem. My Cisco switch is powerd by PoE from the Cisco Firewall. Every time change config of the FW I have to reboot it and because the switch is power via PoE it reboots too --> and the HomeServer has a BOD!

I have just to unplug the switch from power line then the server has a BOD! I dont think that there are any special packets around at that time.

Hi !

I like the comment "*cough cough SIS*" done by you mogwai because I think that's where the problem is. A driver issue...

BR/Daniel

I wanted to drop a line and post saying I have the same problem. I noticed it a couple weeks ago. It happens when I have to power cycle my router of all things! I have a d-link DIR-655.
I found this thread doing a google search and the link was here: http://planken.org/2008/10/25/hp-medias ... y#comments

Lots of folks having this problem.

I have the same issue on my EX470. I analyzed the crash dump via Windbg, and it appears to be a SisGBeXP driver issue forwarding a corruped ethernet packet back to NDIS.

STACK_TEXT:
80894490 f7861f99 0000007c 00000005 8908a278 nt!KeBugCheckEx+0x1b
808944f0 ba932ede 8908a278 80894510 00000001 NDIS!ethFilterDprIndicateReceivePacket+0x6fe
WARNING: Stack unwind information not available. Following frames may be wrong.
80894580 ba92dcd1 00000001 8908a278 89094c68 SiSGbeXP+0x6ede
80894598 f7856466 89094a18 ffdffa40 89094c7c SiSGbeXP+0x1cd1
808945b0 8082f948 89094c7c 89094c68 00000000 NDIS!ndisMDpcX+0x21
80894600 80887dc7 00000000 0000000e 00000000 nt!KiRetireDpcList+0xca
80894604 00000000 0000000e 00000000 00000000 nt!KiIdleLoop+0x2f
(read bottom-up)


I believe the problem is that SisGBeXP is not looking at the IRQ prioritiy level upon the time of their call back to NDIS. KiRetireDpcList drains the DPC pool from either the KiIdleLoop or KiDispatchInterrupt. If it is called from KiIdleLoop, most likely the current IRQL is > DISPATCH_LEVEL. The SiSGbeXP driver should example the IRQL on its callback, if it is running > DISPATCH, then the Ethenet frames should be examined and possibly discarded if its packet frame is empty or corrupted.

-Dave
Sr. Driver Software Engineer

I went ahead and reverted back to the old driver version. I haven't had the problem in a month.

Hate to continue on this old subject but in April, I tossed out my DI-624 WiFi router and replaced it with a Netgear GS108T gigabit switch. Internet connectivity is through my Cradlepoint CTR350 router (with USB EVDO modem), which also handles DHCP.

I used to see frequent crashes (almost daily) and I haven't seen a single glitch since, hardly a coincidence. It may have been a combination of bad packets from the router in combination with a poor driver on the server but, this has solved it for me -- although I spent more money than I intended to.

I don't get it. This has to do with the updated drivers, not a router.