PDA

View Full Version : Raid problems



klapa
11th March 2007, 05:46 PM
Hello all,

I've been having some rather perplexing problems with a set of drives I have configured as RAID1.

The basic symtom of this problem is that these drives seem to be functioning properly - both as individual drives and as a RAID1 set of drives = yet they will always generate an error at system startup if I include them a mounts in /etc/fstab.

If they are included in /etc/fstab, I always receive an error - "superblock indicates drive is size X - but drive is size Y - running fsck recommended".

At this point I must reboot with the rescue CD and remove the mount in fstab to recover.

Some background info:

System is Tyan Tiger 100 motherboard - Intel BX chipset with 2x PII 450 Mhz CPU w/ 512K cache
BIOS vintage ~ 1997
1GB PC100 ECC RAM
system drives = two 10,000 RPM 70 GB SCSI LVDS drives configured as RAID1 - (no problems with these)
data drives = two WD300 ATA IDE drives (these have the problem)

Both the SCSI and IDE controllers are PCI plug-in - the IDE controller is a generic type which has its' own BIOS - this same type of controller works fine in another machine I have with two 40GB MAxtor IDE drives configured as RAID1 system drives - and in fact at one time I had both controllers in this one machine with these two WD300 drives on the second controller - and still had exactly the same problem (only with the WD300 drives).

What is most confusing is that I can mount the raid array from the command line after boot-up with no problems at all - everything seems to work well over days and even weeks of uptime.

I provide some printout here:
================================================== =============
[root@server2 ~]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/md3 57625100 38964752 15685920 72% /
/dev/md0 1019144 43308 923228 5% /boot
tmpfs 517196 0 517196 0% /dev/shm
/dev/md4 3166672 70552 2932664 3% /home
/dev/md2 505508 10577 468832 3% /tmp
/dev/md5 4061504 1825364 2026496 48% /usr
/dev/md1 2030672 194824 1731032 11% /var
/dev/md6 132109124 65018620 60379756 52% /mnt/Nashdom
/dev/md7 156326056 98533924 49851204 67% /mnt/idevol1
[root@server2 ~]# fdisk /dev/hde

The number of cylinders for this disk is set to 36481.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
(e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/hde: 300.0 GB, 300069052416 bytes
255 heads, 63 sectors/track, 36481 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hde1 1 16709 134215011 fd Linux raid autodetect
/dev/hde2 16710 36481 158818590 fd Linux raid autodetect

Command (m for help): q

[root@server2 ~]# fdisk /dev/hdg

The number of cylinders for this disk is set to 36481.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
(e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/hdg: 300.0 GB, 300069052416 bytes
255 heads, 63 sectors/track, 36481 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hdg1 * 1 16709 134215011 fd Linux raid autodetect
/dev/hdg2 16710 36481 158818590 fd Linux raid autodetect

Command (m for help):

================================================== ================
The current RAID state here:
================================================== ================
[root@server2 ~]# mdadm -QD /dev/md6
/dev/md6:
Version : 00.90.03
Creation Time : Wed Feb 28 12:47:59 2007
Raid Level : raid1
Array Size : 134214912 (128.00 GiB 137.44 GB)
Device Size : 134214912 (128.00 GiB 137.44 GB)
Raid Devices : 2
Total Devices : 2
Preferred Minor : 6
Persistence : Superblock is persistent

Update Time : Sun Mar 11 12:31:17 2007
State : clean
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0

UUID : 29ae7363:c11fadee:19c3daa9:39c00030
Events : 0.42

Number Major Minor RaidDevice State
0 34 1 0 active sync /dev/hdg1
1 33 1 1 active sync /dev/hde1
[root@server2 ~]# mdadm -QD /dev/md7
/dev/md7:
Version : 00.90.03
Creation Time : Wed Feb 28 12:49:19 2007
Raid Level : raid1
Array Size : 158818496 (151.46 GiB 162.63 GB)
Device Size : 158818496 (151.46 GiB 162.63 GB)
Raid Devices : 2
Total Devices : 2
Preferred Minor : 7
Persistence : Superblock is persistent

Update Time : Sun Mar 11 10:31:40 2007
State : clean
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0

UUID : 2f579608:af8b4841:0f131152:2040cb78
Events : 0.22

Number Major Minor RaidDevice State
0 34 2 0 active sync /dev/hdg2
1 33 2 1 active sync /dev/hde2
[root@server2 ~]#

================================================== =============

So - everything looks fine from here - yet the sequnce is always the same. If include a line in /etc/fstab to mount these two arrays I will get this failure message a boot time - then I must stop the arrays (releasing all resources), run fsck to fix error (fsck will find errors). Then I will doublecheck with fsck again, and it will find no errors. After this - I can re-create the RAID array, mount it, and everything is just peachy.

I don't include the mount in /etc/fstab, I can re-mount the drives after my next restart with no problems - yet adding the mount in /etc/fstab will once again "damage" the filesystem - and I must repeat the repair sequence.

Anyone have some ideas here?

Dies
12th March 2007, 01:54 PM
Is SELinux active? If so have you tried disabling it?

klapa
12th March 2007, 02:51 PM
Dies,

Thanks for looking - yes - I have disabled SELINUX - as this is really the FIRST thing I do with a new FC6 installation - as it seems the out-of-the-box FC6 installer always sets SELINUX to enforcing without asking for the option during install.

Looking at this problem - I think I have a basic problem with one of those disks - and need to start from scratch.

Dies
12th March 2007, 02:54 PM
It does ask during the initial setup, but it's easy to miss.

That's really the only thing I've seen prevent mounting something at boot that you can mount manually, unless it's a bad option in fstab that you're not using when doing it manually.