topic Re: LH3 Netserver lockup and memory parity error in Netservers

LH3 Netserver lockup and memory parity error

William Harrington_1 — Fri, 11 Nov 2005 12:45:30 GMT

This is the on going saga on one LH3 Netserver file server, we have two.
This machine started about a year ago to take double the time to run backups to tapes accross the network from a unix server. then about 6 months ago it started disallowing user connections after running for about 168 hours. It then degraded to about 96 hours. It then had the top slot and cage two go dead.
We pulled the unit off line and I've attempted to find/correct the problems.
Replaced cage two and all drives are working.
Replaced the I/O system board and reinstalled OS with 2000 and 2003 on various hard drives.
I'm continuing to get a lock up when tring to transfer data accross the network. This machine is at my desk and hooked up with or without a KVM and on the same switch that laser printer and my PC is hooked to.
Because it was hooked up to my switch I could see in one case where the network connection would lag or go amber. The connection now seems good.
The unit will go to sleep overnite and not wake up.
The unit will blue screen with memory parity error.
The unit will freeze and not respond to keyboard or mouse or remote access or ping.

I'm getting to the bottom of the trouble shooting on this unit..

I'm open for suggestions.....Would replacing the processor board be a possible cure?

Re: LH3 Netserver lockup and memory parity error

e4services — Sat, 12 Nov 2005 09:03:12 GMT

I wonder if it could be a simple as a power supply problem? But you also could suspect the CPU and/or the system board itself.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Sat, 12 Nov 2005 15:45:07 GMT

Thanks for the comment.
I've been thinking of pulling the two power supplies, I do have one spare in my other LH3 so will swap on Monday.
Have switched out the CPU from a 450 to a 550 to dual 550's so I'm feel the CPU's are not involved.
So the system board is next, I may just swap them too.
I've not removed the SCSI card for the Hp8000 tape drive. The tape drive is not currently connected so I'll pull tha card on Monday as well.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Mon, 14 Nov 2005 10:09:48 GMT

I pulled the SCSI card today....no improvement.
I replaced the power supplies with a good supply...no improvement.

Found a system board for a reasonable price, will wait for it to show.

Re: LH3 Netserver lockup and memory parity error

Sean T. Craig — Tue, 15 Nov 2005 17:56:06 GMT

Hi William,

Which NIC are you using? Is it the embedded or a PCI card? If it is the embedded, try updating the driver to the latest one from the Intel Website. You can find it at this link:
http://downloadfinder.intel.com/scripts-df-external/Product_Filter.aspx?ProductID=62〈=eng
If that doesn't work, try a PCI NIC and see if it makes any difference.

Best of luck,

Sean T. Craig Sr. C.E.T., M.C.P.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Wed, 16 Nov 2005 08:41:46 GMT

Thanks Sean,
The LH3 doesn't have an integraded NIC. They usally came with two nics, a server NIC in the long PCI slot and Netserver NIC in a conventional slot. The more I test the more I think the problem is involved with the NICS.
I'm starting testing again with the server NIC and Windows2000.

Re: LH3 Netserver lockup and memory parity error

Sean T. Craig — Wed, 16 Nov 2005 10:27:52 GMT

Hi William,

Sorry, should have checked my product specs. I must have been thinking of the LH3000. The driver link in my last post is also the right driver for the HP NIC. Give it a try and let me know.

Good Luck,

Sean T. Craig Sr. C.E.T., M.C.P.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Wed, 16 Nov 2005 11:17:49 GMT

I'll give them a try, Thanks

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Thu, 17 Nov 2005 12:43:11 GMT

Comming back to this question.
Installed the TopTools and can say that the problem is the NIC. If I uninstall the NIC I can do anything want locally and no problem. If the NIC is installed and any serious traffic is generated across the NIC, file transfer, it will run a few MB and then Blue Screen to the parity error. The 3com 905 D or C NICs will not run, they have i amber and 1 green light, the server NIC will run but loses the connection and then reconnects. I can see this lack of connection at the switch as well as on the NIC.
I had a similar problem on a ASUS P4C800 mother board but in that case it was a modem failure. ASUS did replace the board and the problem was cured.
I have replaced the I/O System board which has the PCI slots but that had no effect, the board was refurbished but I doubt that the board would have the identical fault.
I have ordered a used System board.

I would like to know if the LH3 architecture is such that the NIC interface would be routed thru the System board as well as the I/O System Board?

This machine was our network printer and file server for over five years and we have about 20 printers and 130 PC's on the network so it did get a lot of hits. It also ran the back up in the evening of the files as well as the UNIX server data base. It has reason to be tired but like to solve the problem so it can act in a back up and test position.

Does the System Board contain the controls to the NIC interface?

Re: LH3 Netserver lockup and memory parity error

Sean T. Craig — Fri, 18 Nov 2005 13:38:55 GMT

Hi William,

The System Board houses the Processor. Therefore, everything on the I/O board has to go to the system board in some form or another. I'm curious about this Parity Error, though. Can you sent the entire STOP Code? Maybe we can research that a little.

Talk to you soon,

Sean T. Craig Sr. C.E.T. M.C.P.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Fri, 18 Nov 2005 16:55:58 GMT

There is not an actual stop code, the entire message is:
Hardware Malfunction
Callyour Hardware vendor for support
NMI: Parity Check / memory parity error
The system has halted.

At that point it requires a physical power button shutdown and restart. It is very consistant if a NIC is installed that it likes, it doesn't care for the 3COM905 B or C at all.

I've now removed the cards and leaving it powered on wait for the system board to arrive. I've cinvinced myself that the board has just lost the NIC track, we'll know in about 10 days.

Thanks,
Bill

Re: LH3 Netserver lockup and memory parity error

Exeter17 — Fri, 18 Nov 2005 19:14:10 GMT

William

Can you pull the NICs out and try a basic Intel 82557? I have an HP branded one here - pay for shipping and I'll send it to you (it came out of my LH3)

If you are runnning the embedded NR try firmware AND check the lil EDO PARITY SIMM in the back of the box. I had the SIMM go bad and my RIAD performance went to heck after that.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Sat, 19 Nov 2005 17:07:55 GMT

Thanks for the offer on the NIC, will keep it in mind.
I have swapped that small memory card that is located on the upper left corner if the I/O system board between the two LH3's I have and it had no impact.
Not at work now so can not verify it was called a simm but I don't think there are any other small replaceable items that I have not swapped.
Since I have one machine in semi active service I pick and choose my time and parts to swap.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Mon, 21 Nov 2005 08:50:42 GMT

exeter17...if you would e-mail at the address in my profile I would like to take you up on the offer of the NIC.

Re: LH3 Netserver lockup and memory parity error

Paris Finley — Wed, 07 Dec 2005 17:37:41 GMT

WH:
Side note: We have recently come up with the same error (excactly) and symptoms (exactly) on a completely different server configuration, make and model. We think it may be related to a backup tape device managed with Backup Exec (v. 9) that is starting to fail. We see no correlation with the network card yet. I will update you here.--Paris Finley, infosysmgr, Daily Hampshire Gazette

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Thu, 08 Dec 2005 08:59:57 GMT

Thanks for the comment:

This week I received and installed the power share board. This means I've replaced the I/O system board, the system board power share board, one drive cage and controller and the power supplies. If there is not NIC activity the machine runs fine and can handle some light NIC load but during heavy file transfer it still gives the blue screen parity error. I admit these were not new parts and it is possible that these boards have the same problem so I am now planning on swapping parts between the good server and the bad server.
Coincidentaly our first inkling that some was wrong with this server was when the DLT backups started running slow but this was across the network from a UNIX box. It finally lost contact with the DLT and I have not even tried to reinstall the DLT.

This is a project that I will continue to chip away at. Unfortunately we are now putting loads on the other server so I have to pick and choose the best time to swap parts.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Thu, 15 Dec 2005 08:47:54 GMT

My last comment wasn't true. I had two types of errors occuring and assumed they were the same. One was the parity error and one was a freeze up.
Sometimes during the file transfers across the network the machine would freeze, sometimes it would give the parity error.

After replacing the power share board the blue screen parity error did stop however the freeze would still occur, I thought at first that I just happened to hit the freeze and not the parity error but then realized the parity error was not coming back.

This week I pulled the video card and went back to the on board video and the freeze up stopped.

I have subsequently added a Diamond Viper V330 card from an old U-Max PC because the on-board does not give 24bit color resolution and the freeze up has not reoccurred and I've gone back to a Intel pro server NIC card.

Based on the events I recorded it seems that both the I/O board and Power share board were faulty.

I think the problems are cured, but the reality is that very little of the original server is still in place.

Thanks to all, I be closing this thread after the first of the year.

Re: LH3 Netserver lockup and memory parity error

William Harrington_1 — Thu, 22 Dec 2005 13:30:05 GMT

The LH3 Netserver is running normally and stable. Please see my earlier comment