• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1132
  • Last Modified:

inconsistent parity errors on vd strip after restoring image to new machine

We had a SBS server that had some issues, including inconsistent partity on vd strip.

We were sent a new server to replace a bad server. we performed a Windows Bare-Metal Backup using Widows Backup.

On the new server we restored the bare-metal backup and the server came up fine. However running a consistency check on the virtual disks, we still get the parity errors we were getting on the old server

The other symptom we are still getting is the inconsistency check on a 1TB raid 1 drive is taking over 24 hours to complete. Whenever i have performed these in the past, they usually take about 3-4 hours to check 1 TB.

What I want to rule out is the bare-metal restore. Would the restore bring over a bad array which would cause both the parity errors as well as the slow disk checks?
0
Brad Nelson
Asked:
Brad Nelson
  • 4
  • 4
1 Solution
 
DavidCommented:
It can't possibly  be the restore.  When you write a block of data then the RAID controller (or software RAID) writes the necessary parity information at the same time.  

If the hardware is functioning properly, then a restore guarantees there will be no parity errors.

As such, you have a hardware problem.  

Based on your symptoms, it is >99% probability that at least one of your drives as unrecoverable read errors and is doing deep recovery.  One of your drives has to be replaced.  

You could also have problems with both drives.  The only other possible solution (other than you have environmental factors like caustic chemicals, radiation, or unpleasant things that can kill you and your hardware) .. is that your controller is the culprit.  

Suggestion .. turn off all write caching if it is enabled, look at controller logs.  If this was my system, I would shut it down, and attach disks to a non-RAID controller and do some diagnostics and hope it is only one drive, not both.
0
 
Brad NelsonOwnerAuthor Commented:
Dlethe,
That is an excellent reply. And in the same line of thoughts as I. However we have replace the drives already.

 Its a brand new server, with all brand new hardware. I suppose its possible that we had bad drives in the old server, and now bad drives in the new server but the chances seem pretty slim.

I did mention to the customer to put in drives in a non-raid environment to see how they behave. Maybe I'll revisit that idea.

Any other ideas?
0
 
DavidCommented:
What controller & make/model of disks?
0
Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

 
Brad NelsonOwnerAuthor Commented:
Intel Motherboard, Intel RS2WC080 controller. The 1st server was build by a company called Equus and they used incompatible hard drives, which Intel said led to the parity issues.

So once Equus sent us a 2nd server, they provided drives which were on the compatible list. Firmware and drivers for motherboard, back plane are all up to date.
0
 
Brad NelsonOwnerAuthor Commented:
opps forgot to put the disks info.

They are Seagate ST1000NM0001
0
 
DavidCommented:
That controller is based on the LSI 2008 chipset.  It is an excellent controller, but absolutely must have enterprise class drives due to the error recovery firmware settings in the controller.

Any company that builds servers that has OEM relationship with Intel *HAS* to be aware of such constraints. They are inept or can't be trusted.  So humor me, what is make/model of disk.  We do RAID controller & diagnostic work among other things, and have access to stuff from Intel & LSI that may be important.  I can't break any NDA but if there is something I can share then I will.

Alternately, since this is RAID1, then really no reason for a RAID controller.  Windows software RAID1 stack will do load balancing on reads, so you would probably be better off anyway and possibly resolve the problem if you just migrated to software RAID.  Depending on the configurable RAID parameters, you might even see a performance increase.
0
 
DavidCommented:
Those disks are fine, but you could have some grossly incorrect mode page settings.  The disks have extensive non-volatile error logging capability so they can be queried with the software necessary to determine what is root cause.  

There are mode page settings dealing with pre-fetch, retry counts, retry time-outs, write cache, just to name a few, and that controller won't automatically change them to anything optimal.  Off-the-shelf seagate firmware is OK, provided all the hundreds of settings are factory default.  The firmware should be 000n.
0
 
Brad NelsonOwnerAuthor Commented:
drive properties
Here is a screenshot of what the Intel Raid Console see's the drive settings as. I was not able to locate the firmware on these drives. Im guessing there is a seagate tool that could identify that?
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

  • 4
  • 4
Tackle projects and never again get stuck behind a technical roadblock.
Join Now