You are not logged in.
Hi,
I'm using a default kernel 3.6.3 and have a btrfs raid1 (data and metadata) over two hard drives (sda3 and sdb). My system is a subvolume, which works fine. If I remove the first hard drive (which contains grub2), however, and boot using a live cd, I get a kernel panic (image at http://s12.postimage.org/8y0pgvs99/arch … panic.png)
whenever I try to mount the remaining raid1 as degraded
mount -o degraded /dev/sda /mnt(/sda then is the one and only hard disk).
I do this all for testing reasons: I want to simulate a hard drive failure. But right now it just looks like I could not access that data in case of a real failure ;-(
There are some kernel patches
https://patchwork.kernel.org/patch/1267581/
http://www.spinics.net/lists/linux-btrfs/msg17007.html
concerning this issue, which seem to have entered the default kernel, and yet the error persists. (They are contained in the Arch kernel, aren't they?)
Any suggestions?
It's not a bug -- it's a feature!
Offline
In case anyone has the same problems as I did: I found a workaround.
I installed grub on both devices. Since my /boot and / are only subvolumes in the btrfs `device', they are on both physical disks. If the first device fails, I simply boot from the grub of the second disk into the system stored on the second disk (with mount option `degraded') -- and since I mount the btrfs-disk via UUID, I don't have to change /etc/fstab or anything.
It's not a bug -- it's a feature!
Offline