RE: [Evms-devel] EVMS and Drive Linking

[Stian B. B. <st...@ba...>]

[snip]

>> EXT3 disk nr 1:
>> Now I got cocky so I shut down and installed the nr. 2 disk, booted,=20
>> ran fsck.ext3 on it (just to make sure all was dandy), shut down=20
>> again, removed disk nr. 1. Well, now problems occurred. No longer=20
>> could I mount the file system or run fsck.ext3 on it. I read in the=20
>> man page to do a mke2fs -n /dev/evms/ld to find alternate =
superblocks,=20
>> which I found, and then run the command "fsck.ext3 -b "superblock=20
>> number here" /dev/evms/ld". Sadly this did not work, fsck only=20
>> reported "Could this be a zero length partition?". I tried all the=20
>> superblocks I found, but none worked.
>>
>> Cannot find any other way to repair the remains of this partition.=20
>> Maybe if I could remove the placholder I could find some data for =
fsck=20
>> to work with. In my case here it needs to work trough 2,3 GB of=20
>> empty/broken data blocks before finding any data. Maybe it only =
checks=20
>> a little bit?

>Yeah...it would be interesting to see how it reacts to just a blank =
disk in

>that place, instead of the placeholder that always returns I/O errors.

Yup. Just say the word when you have a patch I will be happy to test! ;)

[snip]

>Interesting results. Not very suprising that the filesystems are =
basically=20
>toast if the first disk is lost. I was a bit suprised that ext3 =
survived=20
>losing the middle disk.

Well actually I am surprised that it did not find any data when the =
first
disk is lost even. It should be the case of finding a working superblock =
and
rebuilding the fat table I would guess, but even specifying an alternate
superblock did not produce any results. Perhaps there is something more =
in
the start of the file system it needs. I dunno ... =20

Maybe Drive Linking could accommodate the first 4-5 MB of the first disk =
in
a copy on each disk to go, and when you are unlucky enough to loose the
first disk you could "restore" this first bit to a replace disk and then =
run
a fsck on it. My guess would be that this could work. The new disk would
probably have to be the same size though... or? Well .. it's a thought
anyways.=20

I think I would like to learn some more about file systems, pretty
fascinating. And I think I will mail a copy to the reiserfs guys. They
cannot be very exited about being beaten by ext3 :)

Best regards
Stian B. Barmen