QNAP: Booting with an incomplete mdadm raid (bug 784070)
Hi, on one of my QNAP TS-419P+ NASes running Debian 8.x stable, I started
getting smart errors on one out of the four SATA drives, so I pulled the
drive and swapped it for a brand new one.
The system didn't boot any more.
I found bug 784070 and particularly message 109
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=784070#109
with a hack for
/usr/share/initramfs-tools/scripts/local-top/mdadm
that forces boot to continue with incomplete mdadm arrays.
This worked perfectly; fortunately I could insert the old broken drive, boot
up, log in, change the script, run "update-initramfs -u", shutdown, swap the
drives, reboot and fdisk+mdadm --add the new drive.
I guess without the patch, the NAS is booting to an emergency console, but
there isn't any easily accessible console to interact with.
Is there a chance that something like this can be applied for the next point
update of Debian Jessie stable?
I'm concerned that a local modification to a script like this might be
overwritten at any time, forgotten, and then be left with an unbootable
system the next time something like this happens - especially if the faulting
drive doesn't play ball allowing a quick re-application of the patch.
Thanks.
Reply to: