I need to tie off this thread.
I got my data. I ended up manually (scriptly) merging all the strips.
On later test RAID5 arrays, I removed all the devices. Then I recorded
the configuration of the array from the metadata in the devices. I
cleared the superblocks on the devices and then recreated the array. Of
importance were level, chunk size, and device order. Data offset and
layout were the default so I didn't specify them. The data area was not
changed and, with the same configuration as the last array, was in a
sane format.
************************************************************
Therefore, from my experience, the best way to recover a FAILED array is
to recreate a new array with the same devices and settings.
************************************************************
If anyone less noobish can give a better answer, I welcome it.
PS
Of course, this only works if the RAID devices were manually marked as
failed and really are still good.
On 2017-11-27 07:52 PM, Duane wrote:
OK trying to re-add got me this problem.
# mdadm -A /dev/md0 /dev/sdc2
mdadm: /dev/md0 assembled from 1 drive - not enough to start the array.
# mdadm /dev/md0 --re-add /dev/sdd
mdadm: Cannot get array info for /dev/md0
Looking at the array,
# mdadm -D /dev/md0
/dev/md0:
Version : 1.2
Raid Level : raid0
Total Devices : 1
Persistence : Superblock is persistent
State : inactive
Name : dave:0 (local to host dave)
UUID : f8c0d29d:9b5986a4:050ca914:3a2fb8c8
Events : 51526
Number Major Minor RaidDevice
- 8 34 - /dev/sdc2
I cannot re-add the missing devices.
Doing "mdadm -A /dev/md0 -s" gets me the same array configuration as I
showed in my original post.
Does anyone know why it insists on going to RAID0?
Does anyone know how I can re-add the devices?
Thank you,
Duane
On 2017-11-26 02:11 PM, Wols Lists wrote:
On 26/11/17 17:29, Duane wrote:
You're right I failed and removed device 3, then failed and removed
device 2, which broke my array.
Let's assume I haven't messed things up any more than I already have.
Is there a method to reverse the above operations for device 2 and
device 3?
Thanks, Duane
My raid-fu is mostly theoretical, but what I would try is as follows.
It's assembled the broken array as md0, so
mdadm /dev/md0 --re-add /dev/device2
followed by
mdadm /dev/md0 --re-add /dev/device3
That won't cause any further damage, and may work fine.
Beyond that, I wouldn't like to suggest anything - if it doesn't work
you'll probably need to use things like --force, which could easily
break stuff. But re-add will probably work. And my syntax might not work
- you might need to specify the mode like --grow or --manage, I don't
know.
Suck it and see, at least this won't do any damage, and if it doesn't
work we'll have to wait for further advice.
Cheers,
Wol
On 2017-11-26 05:04 AM, Wols Lists wrote:
On 26/11/17 01:50, Duane wrote:
I have 3 RAID5 devices. When I assemble them, I end up with a RAID0
device.
What is the cause? What is the solution?
All I can think of is the fact that there is only 1 active device.
I had
2 but then manually failed it. I want to reassemble a RAID5 array and
then re-add the second device.
OUCH!
Sorry. You have a 3-device raid-5. You only have 1 working device.
Your
array is well broken.
You can't "reassemble raid5 then readd the second device". You need to
readd the second device in order to get your raid5 back. I'll let
someone else tell you how, but you need a MINIMUM of two devices to
get
your raid working again. Then you need to get your third device added
back otherwise your raid 5 is broken.
DON'T DO ANYTHING WITHOUT ADVICE. I'm sorry, but your message says you
don't understand how raid works, and you are on the verge of
destroying
your array irrevocably. It should be a simple recovery, *provided* you
don't make any mistakes.
Cheers,
Wol
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
null
