I have centos 6.6 with a 4 disk raid 5. One disk is bad.
I removed the bad disk by serial number (it was sdc1).
When i rebooted with a new disk the sd labels changed.
My new disk is now sdb.
Mdadm shows my raid inactive.
Before i do anything else i wanted to know how important are the device names.
Does the same disk need to show up as sdb sdc etcc?
My mdadm conf lists them by name: sdb sdc sdd sde
In my opinion, it would be useful to provide the following information:
Any raid-related messages from /var/log/messages.
Content of configuration file /etc/mdadm.conf
Content of /proc/mdstat
Have you marked the disk sdc1 as faulty via mdadm before replacing hardware?
Are all disk connected to the same controller?
Does controller detect the replaced disk properly, i.e same way as other ones?
Have you connected the new disk to the same controller port as the failed one?
Also please check /var/log/messages to see whether something happens with naming of disks what could case changing of disk names.
If device names has been used instead of UUID then I'd suggest to find out why device names got changed.
