[Wylug-help] RAID etc. revisited...

[Roger]

  Oops!

So I used resize2fs and all was well, so on to the next step - putting 
foot in mouth!

I powered down, removed the old 750G, connected the new 1.5T drives to 
SATA1 & SATA2, hung the 1T drives on SATA 3 & 4 and powered up...

"INSERT BOOT DISK"

Sod it, I thought, the new disks haven't been partitioned and the system 
is looking at the early interfaces to find a boot sector, so powered 
down, swapped the order to the reverse and powered up again.

"INSERT BOOT DISK"

I'd just been called to eat - and I wanted my web-sites back up again - 
so I powered down, grabbed one of the 750G drives and plugged that in 
instead of one of the 1.5T drives, powered back up and the system booted.

After tea, I started to see what was what...  First there was an email 
from mdadm:
---------------------------------------------------

A DegradedArray event had been detected on md device /dev/md.

The /proc/mdstat file currently contains the following:

Personalities : [raid1]
md0 : active raid1 sdb1[1] sda1[0]
       256896 blocks [2/2] [UU]

md1 : active raid1 sdc3[0]
       730266624 blocks [2/1] [U_]
---------------------------------------------------

Note that the 'grown' and resize2fs'd partition has shrunk back to its previous size.  Also
emails sent and received over the last 3 days - mostly those I'd sent to and received from
Wylug-help had - disappeared.

My guess is that I tried to do too much at once, while not making new disks bootable.

The bottom line is that I now have 5 disks physically in the server: 2 new, as yet un-partitioned
1.5TB drives; 2 older 1TB drives not currently mounted in the filing system; plus an "old"
750Gig drive that seems to have no 'knowledge' of anything I've done since planning for the
new drives.  Yet that old drive is the only one currently on line; my "RAID" is currently that
single drive.

To complicate that, before I moved on, I thought to -add the 'faulty' partition (on one of the
1TB drives) back into md1.  My reasoning was that either the fault was a passing glitch, or it
would re-assert itself.

Finally, # fdisk /dev/sdb produces:
----------------------------------------------
Device contains neither a valid DOS partition table, nor Sun, SGI or OSF disklabel
Building a new DOS disklabel. Changes will remain in memory only,
until you decide to write them. After that, of course, the previous
content won't be recoverable.
----------------------------------------------
I pressed q...

Last bits of info: using the other 750GB drive instead also gives an unbootable system; and
I'm running CentOS-5.


Can someone please guide me safely forward?  I think:
1. I shrug my shoulders over the trivial data losses from the last 3 days - but I'm now down
    to a single copy of 500GB of data I can't afford to lose, so can't risk another step back.
2. I think I need to connect the other 1TB drive and see what fdisk says about that before
    moving forward (I'll do that while I wait for guidance)
3. I believe the boot sector must be on a physical drive rather than on the md layer, but if
    I ever knew, I've forgotten how to create one under Linux, other than by a new install

TIA,

Roger


   


______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________