HP MediaSmart server blue screening daily
Every day or so, my HP MediaSmart server crashes with a blue screen and the following event:
Event Source: System Error
Event Category: (102)
Event ID: 1003
Date: 25-Oct-08
Time: 6:37:59 PM
User: N/A
Computer: xxxxxxxx
Description:
Error code 0000007c, parameter1 00000005, parameter2 840c2a38, parameter3 83fccdd8, parameter4 80894510.
This is on a plain-vanilla Windows Home Server without any add-ons installed. It first started occurring October 14, 2008. Perhaps this occurred after one of the Microsoft security patches or one of HP's updates.
I am searching for info on stop code 0x7C. Result so far:
Bug Check 0x7C: BUGCODE_NDIS_DRIVER (Microsoft)
This article points to a NIC problem. The server's NIC is an SiS191 gigabit ethernet card, driver version 2.0.1039.1070. I have downloaded the driver package from SiS (currently the 2.07 package sgl207.zip contains is the newest driver, version 2.0.1039.1110) and I will update the driver tomorrow.
This is problem #3 with this HP MediaSmart Server EX475...
Update October 26, 2008 14:13
I have successfully updated the driver to version 2.0.1039.1110. To do so, I unpacked the ZIP archive sgl207.zip as downloaded in one of the shared folders on the HP server. Then I started the device manager devmgmt.msc in a Terminal Services session (Remote Desktop). Double-click the network card, select the driver tab and click on update driver. Skip Windows' offer to search, and tell it to specify the location. Same on the next screen: select the "don't search" option and then manually go to the unpacked folder.
The correct driver can be found in the Server 2003 folder Srv2003. The actual driver update only takes seconds.
Even though the server did not ask for a reboot, I did this nevertheless.
See what, if any, effect this will have on the blue screen errors...
Update November 1, 2008
After having upgraded the NIC driver, I have not yet seen any STOP errors (blue screen of death, which is a funny term since the HP MediaSmert server is headless).
Update November 9, 2008
The server crashed again with the same STOP code. Also for the first time since the driver update I had connectivity issues again from the console app.
Update February 19, 2009
Since the installation of update HPMSS-1.3-R1.EXE (see January 23, 2009) I haven't experienced a single system crash.
Update February 20, 2009
And not even 24 hours after I thought the problem was solved, the server has crashed again. I give up.
Update February 28, 2009
This morning, the MediaSmart BSOD'd again — only the second time after the HPMSS-1.3-R1.EXE update — and I just happened to be witnessing this unfold.
It started with my D-Link DI-624 router acting up and rebooting. A second later connectivity with the router was restored and I heard the HP MediaSmart rebooting (the fans run at full speed when that happens so it is hard to miss). The event data was exactly the same.
I swapped the D-Link DI-624 router (rev C, firmware 2.76) with another WiFi router I had laying around, a Trendnet TEW-452BRP 'A', firmware 1.6.6. It is not the world's greatest routers (no DHCP reservations, for example) but it was cheap and it works. The network is up and running, we'll see if this makes any difference.
Update March 17, 2009
I have seen two more crashes so the router replacement did not make a difference.
Update May 1, 2009
Almost two weeks ago I have removed the home office quality WiFi router(s) from the network and installed a better Netgear GS108T switch. I started using the Cradlepoint CTR350 router as my main WiFi access point (providing DHCP, routing, etc) I have not yet seen a blue screen on the server and an additional problem I had has also vanished: one of the desktops continuously woke up from hibernation. Plus, now the server and my desktop can actually use Gigabit speed.
Knock on wood, but I may be on to something.
Purchased from NewEgg Canada for about $116. Together
with the $180 CTR350 this was quite a bit more than the
cost of the D-Link DI-624 WiFi router but if this prevents
the MediaSmart server from crashing, I am not complaining.
Update June 23, 2009
Since the installation of the Netgear switch in April, I haven't seen a single anomaly.
Update November 22, 2010
I updated the MediaSmart server OS to version 3. Interestingly, that update still contained the original NIC driver, version 2.0.1039.1070. Since the crashes have not occurred since the switch replacement in April 2009 I have not had the need to update the NIC driver.
Update September 9, 2011
It has almost been 3 years since I wrote this post. The HP MediaSmart server is still running 24/7 with the Netgear switch and the CradlePoint router for WiFi and EVDO access. I have not seen a single blue screen crash since.
Comments
Hi !
I experience the same problem (or at least a similar problem). I have noticed that often I get a BSOD when I reboot my router (DIR-655) or my cable modem. Often I do this while being remotely connected (RDP) to my HP Mediasmart.
This is the latest BSOD I got today:
---
Event Type: Error
Event Source: System Error
Event Category: (102)
Event ID: 1003
Date: 2009-02-21
Time: 14:17:09
User: N/A
Computer: HPSERVER
Description:
Error code 0000007c, parameter1 00000005, parameter2 894b2ab0, parameter3 893d7b50, parameter4 80894510.
---
Checking the crashdump often refers to the SiSGbeXP.sys
driver.
Do you know if your router / modem is having some kind of problem at the time you get the BSOD ?
Maybe time to put some pressure on HP ? :-)
BR/Daniel
I can't say for sure, my router does not seem to retain its logs that long. It's a D-Link DI-624 and these are notorious for rebooting every once and a while (it does have the latest firmware). However, I have never seen it having an effect on the server when it happens, but I can't exclude it either.
I just disabled remote access to the MediaSmart, and disabled any port forwarding on the router; see if that makes any difference.
There does not seem to be any correlation between the crashes and backups.
This morning, the server BSOD'd again, right after my router rebooted. See text.
Hi !
Interesting that you got the BSOD after rebooting the router ! What is the BSOD details ?
/Daniel
Here is the data from the event viewer:
Event Source: System Error
Event Category: (102)
Event ID: 1003
Date: 28-Feb-09
Time: 11:33:06 AM
User: N/A
Computer: HPSERVER
Description:
Error code 0000007c, parameter1 00000005, parameter2 840bdab0, parameter3 83fc82e0, parameter4 80894510.
The only difference are the parameters 2-3:
00000005 840bdab0 83fc82e0 80894510
According to Microsoft:
Parameter 2: address of the miniport block.
Parameter 3: Address of the packet descriptor that the driver uses.
Parameter 4: Address of the packet array that contained this packet descriptor.
A crashing/rebooting router should not cause a server to BSOD, unless there is something wrong with the server hardware. If a different router does not solve it, I was thinking of temporarily using a USB-style network interface for the HP server.
I don't want to make too many changes at once otherwise it will never be clear what the cause is.
Have you experienced any crashes (did you install the HP update as well - it seemed to have decreased the frequency of the crashes)?
Hi !
Well, initially I had crashes but I found out it was a memory issue after I upgraded to 2GB memory. The only issue I've had so far are the BSODs. For me they happen once the router reboots (what I am aware of). My guess is that it must be a driver issue, I mean, the HP server should be able to handle disturbances on the LAN interface.
One thing that could be done I guess is to set the speed on the interface to a fixed one. Either 10 / 100 or 1000. I will try this on both the router (if possible) and the HP server.
/Daniel
Exactly my thought, if a router or switch can bring down a server there is something wrong with the server.
I have exactly the same thing with my HP EX470. I also updated the SiS driver, and have installed all the MS and HP updates.
When the switch reboots or unpluging the network cable the EX470 reboots.
Did anyone of you report this issue to HP ?
Hi !
Happy (or should I be sad) to see more people having the same problem. Yesterday I updated my NIC (SiS191) driver to version 2.0.1039.1110 and I noticed an improvement when it comes to downloading files ! It has improved !
Now what I have to wait and see is if the BSODs will disappear as well...
Are anyone of you running the same driver (2.0.1039.1110) ?
I have not reported this problem to HP yet. Maybe we should ?
BR/Daniel
Hi Daniel!
I also have this issue with driver version 2.0.1039.1110
I created a topic at mediasmartserver.net.
Also, I tried contacting HP but the only thing I saw for Canada was a phone number. I don't have the time to make calls right now, I will do that later.
If any of you have anything new to report, please do so!
The mediasmartserver.net forum topic has not (yet) resulted in the golden tip I was hoping for.
Same HP, same situation. BSOD happens occasionally with nothing but an unknown driver error. But I watched it happen when I reset my router intentionally.
Look forward to hearing some possible solutions. I'm using a linksys router.
Sorry, no solution yet.
The topic at mediasmartserver.net has not resulted in a solution. We are not the only ones, though, and I strongly believe there is an issue with the network card.
The frequent crashes did become a lot less frequent after I swapped routers, but it still happens occasionally. I can only hope it won't happen during some of the extensive file I/O during file duplication...
I was doing a search for gigabit USB NICs (this Linksys one for instance) but I need to find one with the 2003 drivers. This Belkin one seems to have server 2003 drivers.
I have the exact same problem. When ever my router (SpeedTouch 585) is rebooted the EX475 reboots as well with the unextected stop error in event log.
This is really annoying. I too have updated to the latest SIS drivers without anything getting better. Keeps up updated on how your new router is working.
Since I have replaced my router/Cradlepoint combination for the NetGear GS108T/Cradlepoint combination in April, I have not seen any problem.
I have basically merely hidden the symptoms, because the problem lies with the MediaSmart NIC.
Thanks for your comment.
Hi stplanken,
I was just wondering how your restarting MediaSmart has been since your last posting.
It has been a long time since your last update and I know things can go wrong and you forget to go back and update your posts. (Not you specifically - bloggers and people in general!)
Is everything still behaving now that you are on a new switch?
- Brad
Brad: after the last update there was simply nothing new to report. After I replaced that switch I haven't experienced a single BSOD on my MediaSmart server. The only time the server reboots is after any updates that require it. Other than that, it is on 24/7.
The NIC driver update and the switch replacement were really the only changes I made. The BSOD problem is unrelated to the NIC driver. I have successfully updated the HP software to version 3. Interestingly, that update still contains the original driver version 2.0.1039.1070... I haven't had the need to update the driver yet.
Stephan
Thanks for your prompt reply Stephan. It's incredible that over a year since you changed that switch, everything has been top notch.
Currently my computers are all connected to a Netgear DG834PN router and the reboots are getting more and more frequent. I have tried just about everything. I will ask my manager if I can borrow a HP Managed switch this afternoon to see if it makes a difference.
Thanks again Stephan.
- Brad
The reboots seemed to happen whenever there was relatively heavy network traffic. I.e. it never BSOD'd while nobody was home using the server.
In my case, I believe it was due to the generally poor quality of the routers/switches I used (maybe a heat dissipation issue, causing it to crash).
But if you temporarily switch to new switching gear, I would try leaving the "old" switch out of the loop. May not be possible if you need to rely on the WiFi access or any other feature vital to your network...
Have you also upgraded the HP version 3 software? (Not convinced that would solve the BSOD issue but it is a substantial update.) It is not free.
I forgot to mention that I was actually able to reproduce the BSOD by reverting to the old, cheap switches once I had ran with the new setup. Total cost was quite a bit more, but BSOD's are so undesirable that I didn't think twice. Plus, mine weren't gigabit before.
I would agree with you that the problem is caused by large amount of network activity. I can hammer the server with CPU intensive tasks and never seem to run into a problem. My server seems to restart mostly in the early hours of the morning - generally around the times that torrents are scheduled to kick in which of course initiates many simultaneous connections. It's low throughput, but high I/O.
In regards to Version 3 software, my system is an EX490 and came supplied with Version 3. Obviously our hardware is slightly different, but at this stage I have tried everything else apart from a complete system restore and the switch is basically all that is left.
Thanks again for your answers.
- Brad
Sorry to hear that this still is an issue for the new hardware platforms as well -- a bit surprising, actually. I wonder what on-board NIC you have in your EX490.
Device Manager lists it as a Realtek PCIe GBE Family Controller.
What does your server have?
So the NIC is a different make and model altogether. The EX470/475 has a SiS191 gigbit NIC.
In case you are able to (temporarily) take the existing switch out of the loop and it doesn't solve the problem: if you are running any tasks around the time the blue screen crash occurs, see if it is feasible to disable as many functions as possible on the server to exclude the possibility this is a software issue on your server. If this is an office environment this might prove to be difficult, but you have to start somewhere.
Luckily for me this is not an office environment.
The switch has not resolved anything for me - it crashed during the night again, as per usual. A couple of times today as well.
I have found a newer driver from Realtek for the NIC so we will see where that gets me. I'll keep my comments up to date here on the off chance that someone else might benefit from them.
Thanks again for your replies Stephan.
The driver update was one of the first things I tried, it is also the easiest. (See original post.) In my case, it didn't solve the server crashes. And the version 3 software contained the original driver, so I basically downgraded that, but it is still working OK.
Have you tried looking at what you have running at the time of the crash? Typically, after a blue screen crash, the machine reboots, writing an entry in the log file once it has completed the reboot. Maybe you can use the time stamp to see what you had running at the time of the stop.
(Typically, blue screens are hardware/driver issues, but some software installs drivers for one thing or another.)
By the way, maybe you can post the event log entry -- the stop code and parameter values may reveal something.
Just to add to the record. My HomeServer started doing this a couple of weeks ago.
It blue screens randomly while I am at work and it restarts after I restart the Linksys router.
It is funny how it says:
"15.12.2010 - 18:27:16
The system detected that network adapter SiS191 1000/100/10 Ethernet Device was connected to the network, and has initiated normal operation over the network adapter.
---
15.12.2010 - 18:28:23
The previous system shutdown at 18:26:55 on 15.12.2010 was unexpected."
And I've got this one a couple of times:
"16.12.2010 - 07:32:57
Error code 0000007c, parameter1 00000005, parameter2 840948c8, parameter3 83f96490, parameter4 80894510."
Thanks for the info in this post thought, well done.
Thomas: so after your router reboot the server reconnects to the network and then crashes. The stop code parameters are similar.
Interestingly, after I upgraded to the v3 software, the NIC driver was back to the old version (and I still haven't updated that). Judging by the NIC model, is this an EX470/475? What version of the MediaSmart software are you running?
Stephan
Stephen, I am curious if you were using Cat6 cables when you were experiencing these crashes? I am wondering if switching to Cat6 instead of Cat5e would make any difference without buying a new switch. I have already bought 2 different switches and had the problem with both of them.
Sorry to hear that. The quality of cabling is important in the sense that you should always use a good quality. Having said that, I use regular Cat5E. I have never seen any crashes since the router and switch replacement.
The drivers are not the issue either, by the way, as I reverted to the older drivers after the significant V3 software update from HP.
I will just shy away from the cheaper home quality routers and use separate WiFi and hard wired routers/switches. I believe heat dissipation with the cheaper ones may be an issue.
Also, you my want to look at a battery backup for the network equipment, if you don't already have one. Power fluctuations may play a role as well.
By the way, the prices of the routers have dropped significantly. Cradlepoint has also added a few newer models that are a lot less than what I paid for it initially in 2009.
Thanks, Stephen. All the network equipment is on battery backup. I think I am going to see if just going to Cat6 makes any difference before buying yet another switch. If not, I may just replace the router with one that supports gigabit and then I won't need a switch.
That is an option. I feel your pain. Even though I was able to solve the problem, everybody that has this problem is just fighting symptoms, there is an issue with the onboard NIC. I have never explored the option of using a USB NIC, mainly due to throughput, but that is also an option.
I would be surprised if the switch from Cat5e to Cat6 makes a difference here but I hope it works for you. Mind you, with the prices that retailers charge for a simple patch cord, you may be getting close to the price of a gigabit switch.
Maybe you could follow the link earlier in the comments where I point to a forum topic on mediasmartserver.net -- the topic is not attracting a lot of new input due to the age of the server but several other folks have posted their work-arounds or solutions.
Even a new switch may not be the guaranteed solution.
Here is the comment with the link: http://planken.org/2008/10/25/hp-mediasm...
Yeah, I had also considered a USB NIC--in fact, I have an old (non gigabit one) on the server now just to see if it stops the crashes. It's killing me, though, with the slower speed. I can get the Cat6 cables pretty cheap through Monoprice.com and they also have an inexpensive USB gigabit NIC. I'll check out that forum topic.
I'd be interested to know if you find a solution that works for you. The forum post at MediaSmart shows interesting responses.
Just ordered Cat6 and the USB NIC from Monoprice for a total of about $30--will let you know how it goes. That thread was interesting, but didn't have much as far as solutions, except for yours. I'm not ready to chance a new router/switch which may or may not have the same problem, and it's getting urgent because every time the server goes haywire, I get more bad sectors on my disks, which is worrying. Thanks so much for this post. I highly suspected it had something to do with the onboard NIC being faulty but I didn't have the evidence to pin it down to that until I found this page and your thread at the MediaSmart forum.
There used to be a bug in the Windows Home Server OS that caused file corruption when files were left open for a longer period of time when file management tasks were running on the server, but that was fixed with the second update. Did you install any of the updates (the v3 one was offered for a fee, the fist two were free).
How are you seeing the bad sectors on the server, chkdsk? How many bad sectors are you accumulating with each crash? I am not sure what your disk configuration is like, but I would definitely make sure to have backups on external disks and would be hesitant to use the server for anything important. Bad sectors seem to point more towards a hardware issue, or the drive extender.
Try searching the MediaSmartServer.net site for topics covering the bad sectors. Maybe you'll find something interesting there. It returns about ten pages. Maybe if you have anything more specific you can try to narrow the search down a bit.
Also, several topics on the site are referring to an add-in that displays the disk SMART data, that may be interesting if you are thinking of drive failures.
If there are many bad sectors and the number is increasing, I would not open any files on the server from the network and stop the backup service until you have sorted out this bad sector issue.
I am up to date with Power pack 3. I am seeing the bad sectors with the Home Server SMART add-in which I installed a few days ago when I started researching this issue.
I removed the disk with the most bad sectors. It went from 53 to 60 bad sectors during one crash. Of the disks left in my server, the 500G system drive is clean, a 1.5TB drive went from 1 to 3 bad sectors after the last crash, and a 700 GB drive went from 0 to 1 bad sector. I don't have the free space to remove those, but I am going to start backing up the music and photos today. The bad sectors are not enough to qualify for warranty replacement, and they are not detected by chkdsk, but they are reported by SMART.
The Error code 0000007c crash has not occurred since I switched over to a USB NIC yesterday morning--it had been happening daily, but not at a specific time. So I think I am safe for now, but I am minimizing the server use anyway.
Here is the info from the SMART addin about the reallocated sectors:
"Count of reallocated sectors. When the hard drive finds a read/write/verification error, it marks this sector as "reallocated" and transfers data to a special reserved area (spare area). This process is also known as remapping, and "reallocated" sectors are called remaps. This is why, on modern hard disks, "bad sectors" cannot be found while testing the surface – all bad blocks are hidden in reallocated sectors. However, as the number of reallocated sectors increases, the read/write speed tends to decrease. The raw value normally represents a count of the number of bad sectors that have been found and remapped. Thus, the higher the attribute value, the more sectors the drive has had to reallocate."
The system drive is probably not used for backups and data storage, at least that is what I noticed on my server after the v3 upgrade.
Overall, the nbr of bad sectors doesn't strike me as extreme yet, but I wouldn't like it either. One thing I as thinking of is to mount the drive in another computer and run Steve Gibson's Spinrite at level 4. Unfortunately, that is not a free application. I am not sure if there is anything else out there that would be able to accomplish the same thing.
Not one crash or gray server icon since I've been using a USB NIC for a week. Nor has there been any increase in the reallocated sectors on the drives. My gigabit NIC should be here tomorrow.
That is good news. I am glad the USB NIC works for you, although at reduced speed. Personally, I don't really use the media features so for me it wouldn't really be an issue. And any solution that will prevent loss of or corruption of data is good.
Thanks for posting your results.
Just one more update in case others with problems come across this page like I did. This adapter works fine with the EX470: http://www.monoprice.com/products/produc...
(Product ID 5345 from Monoprice.com)
Use the XP/2K 32-bit drivers from here: http://www.asix.com.tw/download.php?sub=...
Filename is AX88178_XP2K_32bit_Driver_v1.4.3.64_WHQL.zip
Hope it helps someone!
Ah, thanks for the links, I forgot to ask what you used, glad you posted it. (I edited your comment to properly show the links.) I am wondering if you will actually notice the difference when compared to the on-board gigabit NIC under normal circumstances.
And thanks for posting your experience here, I am glad you did.
- Stephan
I just transferred a bunch of new music and it seems just as fast to me.
That is good news. I use the server typically as a file server (with its nice redundancy in the form of the drive extender) and for backups and typically this requires surprisingly little bandwidth. I'm glad you found a good work-around!
Thanks again for sharing your experience!