18 months of trouble with Small Business Server
- From: Fulco Stiva <fulco@xxxxxxxxxxxxx>
- Date: Fri, 26 Jun 2009 19:19:20 +0200
18 months of trouble with Small Business Server
In December 2007 I bough an Intel Server with MS SBS 2003.
I wanted a care-free Server.
So I bought all hardware form one manufacturer (Intel), Enterprise quality harddisk from Seagate, and Kingston RAM.
This would give me a High Quality hardware platform.
The Software I planned to use was SBS 2003 SP1.
All hardware (drivers) and software where listed as being compatible.
This was a total clean installation: NOT a migration!
But after installation the problems started:
I was unable to get an ISDN (Problem 1) solution to work.
At first I used my old server as production server, so business could continue.
Then after some weeks I started using the new system (production).
But soon after, the other problems (Problem 2) became more visible.
Sometimes writing to disk became exceptionally slow (< 2 Mb/s).
Sometimes disks disappeared from the system (including the System disk).
Restarts failed due to missing System drive.
There where: System Hang, Crash without BSOD.
Most crashes (and Restarts) resulted in dead system. I had to pull the power-plug. And there was never information in the System Event log.
After 8 months of crashes I replaced the Intel SRCSAS18E RAID Controller for an Adaptec 5805.
This RAID Controller had more harddisk on it compatibility list (including the Seagate disks I used).
But this didn’t resolved the crashes.
During this time I also replaced most replaceable hardware.
Then SBS 2008 SP1 became available:
My (stupid) thoughts where: based on Windows 2008 SP1, 100% compatible (on paper), maybe this solves every issue.
NO!
I did several clean installations (on new Hitachi SAS disks), and made an installation protocol for every step.
I could install SBS 2008 without adding any drivers: every necessary driver was available on the MS SBS 2008 DVD (and compatible).
But there was no difference using the drivers on the DVD or the most recent drivers from Intel/Adaptec.
After installing the 5805 Management Software there where even more problems (Hang and Crash).
I found: Disabling the Adaptec Agent (Service) stop the BSOD. (BSOD: STOP 0x...D1, 0x,...34C, DRIVER_IRQL_NOT_LESS_OR_EQUAL, tscpip.sys)
I contacted Adaptec: they where unable to reproduce this problem.
Then I did 3 more Clean (and Format) installations: always the same result.
Eventually the slow writes where solved (Solution 1)
Last week the “Drives getting lost” issue seems solved with the new HBA firmware (Solution 2). I am not totally convinced yet, but until now no more failing Restarts.
In total I spend more then ??? hours, did at-least 10 full (uncountable partial installation) clean installations (SBS 2003 and SBS 2008), and spend almost twice the amount of money on the server:
But still there are a lot of crashes.
Yesterday I made a mistake: after updating Adaptec drivers and software, I forgot to Disable the Adaptec Agent, which is automatically installed and started after updating the Adaptec Storage Management Software.
This resulted in the usual crash (BSOD: STOP 0x...D1, 0x,...34C, DRIVER_IRQL_NOT_LESS_OR_EQUAL, tscpip.sys).
This time the crash damaged the boot sector (?): chckdsk C: "second NTFS boot sector unwritable" and Windows Backup fails every time.
ERROR: Chkdsk C:\ "Second NTFS boot sector unwriteable"
ERROR: Backup fails every time with Shadow Copy errors
Please help me solve these two error.
Fulco
The Problems:
Problem 1) ISDN:
I tried to install an ISDN-card (PrimuX S0 NT).
Even with extended support from the manufacturer, I didn't succeed get the product to work with S5000PSL.
Also we tried the PrimuX USB, which didn't work either( there where USB-Resets disconnecting the device).
Problem 2) Hang (need to pull the plug), Crashes, Slow disk copy, Drives getting lost:
Most of the time after crash one or more Drive Cages is missing.
Most of the time No crash information in Windows.
Slow and unreliable harddisk writes.
Restart in OS: very often one or more Drive Cages where missing (after restart NOT A CRASH).
Need to pull the power-cables. Wait a few minutes. Connect cables, boot and drives are back.
AND: almost every crash, hang there was NO BSOD, and NO debug information like Minidump or MEMEORY.DMP.
AMD: the only System Event was: last Shutdown was unexpected, no event what happend.
Problem 3) Intel Management:
RMM2 (Intel Remote Management Hardware): no remote console posible (Vista/MacOS/XP, can login to RMM2), SEL log problems
Intel Management: software interferes with windows OS: unable to install and use Intel Management
Due to the other issue I even started looking into this.
The System:
OS:
Windows Small Business Server 2008 SP1 (64 bit)
Windows Small Business Server 2003 R2 SP2 (32 bit) (replaced by SBS 2008)
Hardware (Intel):
Chassis: SC5400LX
Motherboard: S5000PSL
RMM: RMM2 (Intel Remote Management hardware)
Drive cages: AXX4DRV3GEXP and AXX6DRV3GEXP (this are SAS/SATA HBAs with build-in Expander)
RAM: Kingston
Harddisks: Seagate, Hitachi and Western Digital
Firmware/Drivers:
all components have latest versions, updated when a final version when they become available
The trouble shooting:
So far I tried:
- Replaced most Harddisks
- Replaced Drive Cages's EXP4 and EXP6 (Intel HBAs)
- Replaced RAID controller (Intel SRCSAS18E -> replaced for Adaptec 5805)
- Replaced OS (new clean installations): before SBS 2003 -> now SBS 2008
- Replaced RAM
Primux Support:
Lots of support. They send me several PCI boards to test. They even made a hardware modification.
Intel Support:
Dozens of emails.
Adaptec Support:
Dozens of emails and Adaptec send me a new 5805.
Some problems I solved:
Solutions 1):
Seagate Barracuda ES / Barracuda ES.2 problems: Long before Intel and Adaptec published articles describing problems with these disk. Even if the Barracuda had the Firmware stated in the Harddisk Compatiblility list of the RAID Controller, there still where issue. The Harddisk Compatibility List for the RAID controllers was updated more then one time. I spent a large amount of time solving the problem. I replaced many disk (Hitachi and Western Digital). Then new firmware became available. This solved partially the issue. A second/third firmware update solved these problems. I will never buy another Seagate drive!
Solutions 2):
For a few weeks a firmware update for the Intel HBAs became available. The update to version 2.12 seems to solve ”the missing drives after restart” (genuine Restarts and Crashes) issue. In the release notes: “Issues Fixed: Expander enclose (HBS) may go offline causing a RAID failure.” Thanks Intel I reported this Q1 2008!
.
- Follow-Ups:
- Re: 18 months of trouble with Small Business Server
- From: Leythos
- Re: 18 months of trouble with Small Business Server
- Prev by Date: Re: Is it possible to do a Repair Install on Dynamic Mirror Disks?
- Next by Date: Re: RWW SBS 2008 connection issues
- Previous by thread: Is it possible to do a Repair Install on Dynamic Mirror Disks?
- Next by thread: Re: 18 months of trouble with Small Business Server
- Index(es):
Relevant Pages
|
