🚀 go-pugleaf

John W. Fuerst wrote:

> Hi,
> 
> How can I prevent the RAID1 for root from doing a resync with each boot?
> Shouldn't the RAID be clean after a regular shutdown?
> I also did a 'fsck' on /dev/md0 from the rescue system. But next time
> after the system was shutdown, when /dev/md0 was the root device (rw),
> it would resync again with the next boot.
> 
> /dev/md1 is perfectly fine and resyncs not too often.
> 
> Some advice, please.
> 
> 
> The setup:
> I have setup my system to boot with GRUB from a harddisk device and have
> the root filesystem on a RAID1 device. /home is on another RAID1 device.
> The two harddisks are on seperate controllers.
> Everything works fine exept the fact that '/dev/md0' is resyncing
> everytime after boot. Is this normal or can you show a way to avoid that.
> 
> Info about the configuraton:
> 
> SuSE8.2, kernel 2.4.20-4GB (compiled with reiserfs.o and raid1.0
> builtin), raidtools V1.00.3-24
> 
I would try a more recent kernel. Using softraid by myself here, I noticed
that unnecessary resyncs have gone away completely at least since I run
2.4.24 or 2.6.x.
.....
> <6>md: hde2's event counter: 00000025
> <6>md: hda2's event counter: 00000025
> <3>md: md0: raid array is not clean -- starting background reconstruction
> <6>md: RAID level 1 does not need chunksize! Continuing anyway.
> <6>md0: max total readahead window set to 508k
> <6>md0: 1 data-disks, max readahead per data-disk: 508k
> <6>raid1: device hde2 operational as mirror 0
> <6>raid1: device hda2 operational as mirror 1
> <4>raid1: raid set md0 not clean; reconstructing mirrors
> <6>raid1: raid set md0 active with 2 out of 2 mirrors
> <6>md: updating md0 RAID superblock on device
> <6>md: hde2 [events: 00000026]<6>(write) hde2's sb offset: 45865408
> <6>md: syncing RAID array md0
> <6>md: minimum _guaranteed_ reconstruction speed: 100 KB/sec/disc.
> <6>md: using maximum available idle IO bandwith (but not more than
> 100000 KB/sec
> ) for reconstruction.
> <6>md: using 508k window, over a total of 45865408 blocks.
> <6>md: hda2 [events: 00000026]<6>(write) hda2's sb offset: 45865472
> <6> [events: 00000000]
> <3>md: invalid raid superblock magic on md0
> <4>md: md0 has invalid sb, not importing!

Appears the raidtools got something wrong here ...

> <4>md: no nested md device found
> <6>md: ... autorun DONE.

I would suggest you do 
- boot from - best choice - a knoppix 3.3 or 3.4 cd and
- copy raidtab from your mounted etc partition (you may as well mount only
one disk of the raid but r/o) to /etc in ramdisk
- unmount your distributions root again
- raidstop /dev/md0
- fsck -y /dev/hde2
- dd if=/dev/hde2 of=/dev/hda2 bs=64k
Make absolutely sure the command completes without error, except it
complains about the last block not being written completely.
mkraid --(really-)force --dangerous-no-resync /dev/md0
Well, that last command sounds a bit "dangerous" indeed, but for me it did
always work fine. Nothing more superfluous to tell but "backups are a
must-have" ;-).

-- 
Longhorn error#4711: TCPA / NGSCB VIOLATION: Microsoft optical mouse 
detected penguin patterns on mousepad. Partition scan in progress
�to�remove�offending�incompatible�products.��Reactivate�your�MS�software. 
Linux woodpecker.homnet.at 2.6.4-1.tmb.2mdk�[LinuxCounter#295241]