Re: system upgrade reordered drives, confused software raid-1

On Tue, Jan 20, 2009 at 6:14 AM, Justin Piszcz [off-list ref] wrote:

On Mon, 19 Jan 2009, Troy Cauble wrote:

quoted

I recently upgraded an Ubuntu box and noticed that my drives got reordered
(possibly by a SATA driver change) and my RAID-1 is degraded.

Originally...
My (PATA) OS drive had some other partitions and *was* known as /dev/sda.
My (SATA) raid-1 /home *was* made from /dev/sdb and /dev/sdc.

Now my OS drive is sdc with side effects like mounting some sdc partitions
on my old sda mount points. (not a big deal)

The real problem is the RAID-1. I think the following 3 queries say:
1) the raid is running with SDA, with the other drive removed.
2) SDA is clean and part of md0
3) SDB is "clean", but part of a raid that doesn't exist.


Quote:
mastershake:~> sudo mdadm --detail /dev/md0
/dev/md0:
Version : 00.90.03
Creation Time : Sun Jan 2 12:53:01 2005
Raid Level : raid1
Array Size : 244195904 (232.88 GiB 250.06 GB)
Used Dev Size : 244195904 (232.88 GiB 250.06 GB)
Raid Devices : 2
Total Devices : 1
Preferred Minor : 0
Persistence : Superblock is persistent

Update Time : Sun Jan 18 23:33:56 2009
State : clean, degraded
Active Devices : 1
Working Devices : 1
Failed Devices : 0
Spare Devices : 0

UUID : 61645020:223a69dc:12d77363:0c0f047d
Events : 0.5125062

Number Major Minor RaidDevice State
0 8 1 0 active sync /dev/sda1
1 0 0 1 removed


Quote:
mastershake:~> sudo mdadm --examine /dev/sda1
/dev/sda1:
Magic : a92b4efc
Version : 00.90.00
UUID : 61645020:223a69dc:12d77363:0c0f047d
Creation Time : Sun Jan 2 12:53:01 2005
Raid Level : raid1
Used Dev Size : 244195904 (232.88 GiB 250.06 GB)
Array Size : 244195904 (232.88 GiB 250.06 GB)
Raid Devices : 2
Total Devices : 1
Preferred Minor : 0

Update Time : Sun Jan 18 23:34:36 2009
State : clean
Active Devices : 1
Working Devices : 1
Failed Devices : 1
Spare Devices : 0
Checksum : e627430d - correct
Events : 0.5125064


Number Major Minor RaidDevice State
this 0 8 1 0 active sync /dev/sda1

0 0 8 1 0 active sync /dev/sda1
1 1 0 0 1 faulty removed
Quote:


mastershake:~> sudo mdadm --examine /dev/sdb1
/dev/sdb1:
Magic : a92b4efc
Version : 00.90.00
UUID : 61645020:223a69dc:12d77363:0c0f047d
Creation Time : Sun Jan 2 12:53:01 2005
Raid Level : raid1
Used Dev Size : 244195904 (232.88 GiB 250.06 GB)
Array Size : 244195904 (232.88 GiB 250.06 GB)
Raid Devices : 2
Total Devices : 2
Preferred Minor : 0

Update Time : Tue Sep 16 07:38:13 2008
State : clean
Active Devices : 2
Working Devices : 2
Failed Devices : 0
Spare Devices : 0
Checksum : e5808830 - correct
Events : 0.5048904


Number Major Minor RaidDevice State
this 1 8 33 1 active sync /dev/sdc1

0 0 8 17 0 active sync /dev/sdb1
1 1 8 33 1 active sync /dev/sdc1


I've touched files on the degraded md0, so sdb is out-of-date.

So what's the safest way to fix this? I've never had to rebuild before.

Do I need to fail, remove, then add sdb?  Can I
  mdadm -f /dev/md0 /dev/sdb1
if md0 doesn't thinks sdb is part of it?

Or does "removed" in the md0 output mean I don't need to
fail and remove, just add?

Or do I need to "assemble"?  I really don't understand when to use what.

Thanks,
-troy

1. mdadm /dev/md0 --fail /dev/sdb1
2. mdadm /dev/md0 -r /dev/sdb1
3. sfdisk -d /dev/sda | sfdisk /dev/sdb
4. mdadm /dev/md0 -a /dev/sdb1
5. mdadm --examine --scan >> /etc/mdadm/mdadm.conf

Justin.

Thanks very much!

If I understand the documentation, this is essentially treating
sdb1 as a new drive and sda1 will be copied too it.

I'd also like to understand if/when --assemble might be appropriate for
these scenarios.  The man page says Assemble means

"Assemble  the  components  of a previously created array into an
              active array."

My two drives were part of a previous array that are no longer
associated for some reason.  I *assume* that sdb1 still has a
good file system, just slightly out of date files, since I used the
degraded array.

Is it that assemble wouldn't know which version is correct?
Is it that assemble isn't used for recovery scenarios?
Is it that sdb1 might be worse than I assume?

Thanks,
-troy