This post is more than 5 years old
23 Posts
0
32550
March 24th, 2010 22:00
CRC errors on RAID1
Hi,
I have a PE1900 Microsoft Small Business Server 2003 box with a PERC 4/SC RAID1 logical drive. I just know enought to get myself in trouble so here is where I am at before I do anything else.
The logical raid drive has 2 volumes as follows: C as the system volume, and E as the DATA volume. I DO NOT have a last known good NTbackup of this server. The last known good NTbackup job was over written by a failed backup job. This is due to cyclic redundancy check errors on many system files. I've only used the integrated NT backup software and no is no other thirdparty software installed at this point in time.
Here is what I think I must do:
1. I must backup this entire drive before I touch it. If someone can suggest a third party image software or backup software program for specifically the SBS OS please let me know what works best for this situation.
2. In order to determine if this is a hardware issue or software issue are there any other tools that Dell provides to determine my this next objective or do I just do the chkdsk /f command?
3. I also realize that I cannot complete the chkdsk /f action while the drives are in a RAID 1 array due to the fact that this will break the RAID1. How can I break the RAID 1 array and then complete my hardware or software troubleshooting procedures as outlined?
4. If I 've missed something please feel free to correct me. I rarely run into this type of situation and don't want to screw up. Humbly thank everyone for any assitance in advance.
Joe
candulj
23 Posts
1
March 25th, 2010 07:00
Hi,
Thank you for your input but please correct me if I am wrong. From my understanding the windows chkdsk utility is designed to correct filesystem errors on hard disks. They are not and can not correct non-filesystem errors on hard disks. This means that it WILL have an effect on the current RAID array. Should I not backup the RAID and run a consistency check on the RAID and if successful then gracefully shut the server down and then break the RAID followed by the chkdsk command? Again thanks for your input.
theflash1932
11 Legend
•
16.3K Posts
1
March 25th, 2010 08:00
Yes, the chkdsk will only fix (and CAN only fix) file system errors, so we have the Consistency Check, which will only fix errors within the logical drive managed by the RAID controller. The RAID controller creates/manages a logical drive, made up of however many disks, in whatever configuration it was told to use, then hands the OS a single unit of storage space. Windows doesn't know and doesn't care how many disks make up the array it sits on or what kind of array it is, be it RAID 0, 1, 5, 50, etc - or a single drive connected to an IDE slot. Dell's diagnostics can check individual disks connected to a RAID controller for hardware errors, a consistency check will check the logical drive (array) for errors, and chkdsk will check the file system for errors. They should be done in that order, but because chkdsk (Windows) is unaware of the underlying structure of the logical drive it sits on, only the file system it resides on will be affected. This isn't just theory ... this is hundreds of chkdsk's run on a variety of hardware, RAID levels (degraded and healthy), and for a variety of problems :)
Now, we can throw all this out the window if you are running a Windows-managed (Disk Management) Mirror/RAID 1. This applies to hardware RAID only.
theflash1932
11 Legend
•
16.3K Posts
0
March 24th, 2010 22:00
Dell does have 32-bit diagnostics that you can run to determine if there is a hardware failure somewhere (hard drives).
If you are running hardware RAID (PERC 4/SC - in a 1900??), then CHKDSK will have no effect on the RAID, as Windows is oblivious to the hardware RAID it sits on. I'm not familiar enough with software RAID (RAID managed by Windows Disk Management) to know for sure on that, but I cannot see how that would be a problem either. You may however want to run a Consistency Check on the array before doing your CHKDSK.
pcmeiners
6 Operator
•
1.8K Posts
0
March 25th, 2010 07:00
The only way I know of a crc occurring due to a raid issue is if you had raid card memory errors or maybe a firmware bug. Chkdsk /f will not break a raid, with the possible exception of a raid created within the Windows OS via dynamic disks; if you have very serious issues with the OS then chkdsk could stop Windows in it's tracks..but the OS would already be ready to fall apart anyway.
"How can I break the RAID 1 array and then complete my hardware or software troubleshooting procedures as outlined?"
Breaking a raid to test individual disks does not accomplish anything. Even if you did test individual disk on a standard SCSI adapter, the test are useless, as raid adapters check disks differently then if they are on SCSI HBAs. Years ago I would take SCSI drives which failed on a raid,place them on SCSI adapters for testing, many would pass continuous testing over three weeks, with not errors, using multiple drive testing softwares/hardware tester. Place back on a raid adapter they would fail again, but used off a standard SCSI interface lasted for years without issues.
As far as imaging, hit or miss as to whether the image will work, as images or cloning does not like drives which have errors on them.