my server (Debian Wheezy) has a failed disk. I have sda and sdb and according to the support of my hoster, sdb needs a replacement.
I send them this information with the mdstat: http://pastebin.com/GrzECixk
So they took the server down, checked the disks, said sdb is defect and I need to manually stop the software raid. I tried to use my hosters tutorial for that: http://wiki.hetzner.de/index.php/Festplattenaustausch_im_Software-RAID/en (I am not sure if that is the English version)
But the catch is that mdstat has changed since the technician is done:
# cat /proc/mdstat Personalities : [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] md125 : active raid1 sdb1 2096064 blocks [2/1] [_U] md126 : active (auto-read-only) raid1 sdb2 524224 blocks [2/1] [_U] md127 : active (auto-read-only) raid1 sdb3 1462516672 blocks [2/1] [_U] md2 : active raid1 sda3 1462516672 blocks [2/1] [U_] md1 : active raid1 sda2 524224 blocks [2/1] [U_] md0 : active raid1 sda1 2096064 blocks [2/1] [U_]
I tried removing md125, it said the device or resource is busy, so I set it to failed and tried again, still busy.
So what do I need to do now? How can I disable sdb first and give the technician the go to replace the disk?
Additional mdadm -E output for each sda partition: pastebin.com/J8ytbVgt
for sdb partitions: pastebin.com/RFyDygZj