On 02/19/2017 08:49 PM, Martin Bosner wrote:

I am running a software raid6 with 36 x 3TB disks (sda to sdaj). All
disks have one partition (gpt, 100%, primary, raid on) and i am using
btrfs on top of the raid.

Last week one of the disks failed and was unrecoverable. I replaced the
disk (sdk) with a new one and the resync process started.
At around 80% recovery two further disks failed and the recovery process
was stopped. That failed disks are sdm and sdh.

All other disks seem to be fine and I was about the use the "mdadm
--create" command when i remembered the lines
"You have been warned! It's better to send an email to the linux-raid
mailing list with detailed information"

So here i am for an advice how to continue.

More information, please.  Paste inline, untrimmed, in your reply with
line wrapping disabled.  Plain text only.  Use multiple mails if needed.
List limit is ~100k IIRC.

# dmesg

# for x in /dev/sd[a-z] /dev/sda[a-j] ; do echo mdadm -E ${x}1 ; smartctl -iA -l scterc $x ; done

Phil