Bug#784070: mdadm Software RAID1 with GPT on Debian 8.0.0 amd64 - Does not mount/boot on disk removal

[Robert.K.]

I am also hit by this bug in Debian Jessie 8. I got it in a virtualized
Virtualbox machine with one virtual disk (VDI file) and one disk attached
through USB where the root was located on a md device.

The first error message that appeared was: Gave up waiting for root device
It is not mentioned in this bug report so if people search for the message
they won't find it, until now hopefully.

I used the workaround described on serverfault which works for now. Adding
bootdegraded=1 and such to the kernel command line does not work.

I don't know if it has to do with mdadm but before trying to  detach one of
the disks (to test an eventual disk failure) I in fact had problems with
mdadm not waiting for the slower USB disk to become available. I solved
this by adding rootdelay=1 to /etc/default/grub

I changed this line:
GRUB_CMDLINE_LINUX_DEFAULT="quiet"

To this:
GRUB_CMDLINE_LINUX_DEFAULT="quiet rootdelay=15"

And then I ran update-grub.

I consider this as a critical bug. RAID1 systems which should be more fail
tolerant may currently not boot in Debian Jessie 8 if one of the disks are
out of order.

r
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-mdadm-devel/attachments/20150610/83d6f3e4/attachment.html>