-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Alvin,
Well, after some backing up, i decided to take a risk and put a new root
drive in the machine with RedHat 6.0 and Raidtools 0.90.
Spent a good 15mins looking for ckraid which obviously didnt exist, and
digesting how the kernel now auto-runs/repairs the raid arrays.
Firstly i ran an mkraid --upgrade /dev/md0, which seemed to upgrade the array
just fine. On later inspection i noticed i had an old /etc/raidtab file
installed, so i fixed up the raidtab to how it should be, and spent the next
30mins trying to work out how to get the kernel to 'ckraid' the array.
After reading /usr/doc/raidtools* i noted that the members of the RAID array
needed a file type of '0xFD' so i proceeded to set the 'type' on the drives to
'fd' using linux's fdisk program (naturally after making sure the array was
down).
Still no success. At this time i was getting close to re-creating the array &
reformatting.
A last ditch effort was to run 'mkraid --force-resync /dev/md0', apart from a
"WARNING: All data will be lost" message it seemed to update the superblocks
just fine.
A reboot later revealed success, /dev/md0 was started, allbeit in degraded mode
(probably due to the superblocks being messed with) and all my data intact.
Many hours later after running the machine 'live' the background repair process
finished, and i seem to have my RAID array back.
Here is whats in my /proc/mdstat:
Personalities : [raid5]
read_ahead 1024 sectors
md0 : active raid5 sdd1[3] sdc1[2] sdb1[1] sda1[0] 26651520 blocks level 5, 4k
chunk, algorithm 2 [4/4] [UUUU]
unused devices: <none>
My current /etc/raidtab:
# Sample raid-5 configuration
raiddev /dev/md0
raid-level 5
nr-raid-disks 4
chunk-size 4
persistent-superblock 1
# Parity placement algorithm
#parity-algorithm left-asymmetric
parity-algorithm left-symmetric
#parity-algorithm right-asymmetric
#parity-algorithm right-symmetric
# Spare disks for hot reconstruction (not supported yet)
#nr-spare-disks 0
device /dev/sda1
raid-disk 0
device /dev/sdb1
raid-disk 1
device /dev/sdc1
raid-disk 2
device /dev/sdd1
raid-disk 3
Hopefully this information can help others performing similar upgrades, as to
things to watch out for.
Regards,
Chris.
On 24-Oct-99 Alvin Oga wrote:
> hi ya chris
>
>> I have a RAID array created using raidtools-0.41 on a 2.0.36 kernel (RedHat
>> 5.0
>> i think), which is running Linux software RAID-5 just fine.
>>
>> I am updating the software to RedHat v6.0 soon which uses raidtools-0.90 and
>> a
>> 2.2.5 kernel.
>>
>> Anyone see any potential problems or have war stories to trade?
>
> I did that exact upgrade about 6 months ago....
>
> I had to use the old mdadd/mdrun commands from the old raidtools...
>
> Do not use the new raidtools if you value your old data
> in upgrading from redhat-5.x to redhat-6.0.....
> ( lost my backup disks during my upgrade tests... not a big issue )
> - note redhat-6.x also does additional reverse dns checks
> and needs ver=2 for talking/mounting the older sun boxes too
>
> am currently upgrading to redhat-6.1..... and so far... those older
> superblocks
> still seems to be okay with the new redhat-6.1 (2.2.10) /etc/raidtab
> format/tools...
> - am using hardware raid ( three 64Gb boxes ) + raid0 for
mirroring/backups
> w/ many 15Gb ide drives spanning a few servers ( daily, weekly, monthly
)...
>
> have fun "raiding"
> alvin
> http://www.linux-consulting.com/Raid ( my collection of raid notes )
>
- -- "Failure is not an option, it comes bundled with your Microsoft product."
- ----------------------------------
Chris Keladis
TPG Internet System Administrator
E-Mail: Chris Keladis <[EMAIL PROTECTED]>
Local Date: 25-Oct-99
Local Time: 22:26:38
- ----------------------------------
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 5.0i for non-commercial use
Charset: noconv
iQA/AwUBOBRSg326Uwt427DhEQLsigCeOPDWdQgxpvS3g712B1y0RXFsgfwAoOQP
VpkubF/+oLDRjHEYQxKez2mE
=9w1o
-----END PGP SIGNATURE-----
