array degraded after reboot

array degraded after reboot

[Jeremy Huddleston]

[-- Attachment #1: Type: text/plain, Size: 748 bytes --]

I wanted to remove /dev/hdk6 from my array and replace it with a new 
drive at /dev/hdj6 (I know about the efficiency problems involved with 
this, but trust me that this is what I want to do for some other 
reasons), so I did the following:

mdadm -f /dev/md6 /dev/hdk6
mdadm -r /dev/md6 /dev/hdk6
mdadm -a /dev/md6 /dev/hdj6

The RAID reonsctucted itself, and it was working fine, so I rebooted the 
machine and removed the /dev/hdk hard drive from my computer.  When 
linux started up again, I checked /proc/mdstat, and it said that the 
RAID was running in degraded mode without /dev/hdj6.

I did the process again and stored information at each step (you can see 
this in the tarball).

I am using mdadm 1.2.0.

Thanks for the help.

--jeremy

[-- Attachment #2: mdebug.tar.gz --]
[-- Type: application/x-gzip, Size: 1921 bytes --]

Re: array degraded after reboot

[Neil Brown]

On Thursday May 1, jeremyhu@uclink4.berkeley.edu wrote:
> I wanted to remove /dev/hdk6 from my array and replace it with a new 
> drive at /dev/hdj6 (I know about the efficiency problems involved with 
> this, but trust me that this is what I want to do for some other 
> reasons), so I did the following:
> 
> mdadm -f /dev/md6 /dev/hdk6
> mdadm -r /dev/md6 /dev/hdk6
> mdadm -a /dev/md6 /dev/hdj6
> 
> The RAID reonsctucted itself, and it was working fine, so I rebooted the 
> machine and removed the /dev/hdk hard drive from my computer.  When 
> linux started up again, I checked /proc/mdstat, and it said that the 
> RAID was running in degraded mode without /dev/hdj6.

At a guess, I would say that you are depending on in-kernel
auto-detect to assemble your arrays, and you haven't set the partition
type of hdj6 to "Linux Raid Autodetect"

NeilBrown