[pkg-cryptsetup-devel] Bug#933059: Debian Buster with encrypted root on degraded raid1 (md-raid)
guilhem at debian.org
Wed Dec 30 23:35:38 GMT 2020
Control: tag -1 pending
Thanks all for the patches and discussion, and sorry for not chiming in
earlier in the release cycle.
I now merged in Guilherme's patch modulo some minor fixes. My first
reaction was this this was an “abuse” of initramfs-tools(7)'s interface
since it clearly state that passed local-top stage the root device node
is expected to be present (encouraging us to call the local-block/mdadm
from local-top/cryptroot, like we're already doing for local-block/lvm2),
however it also says later that if local-top fails to create the root
device node the local-block scripts will be called periodically.
So in the end I like Guilherme's approach to give up early at local-top
stage and let the local-block caller take care of the retry and polling
logic. In fact on closer look it appears initramfs-tools(7) is now
discouraging to do any kind of polling:
This should allow further simplifying the boot script: no need to loop
up to $ROOTDELAY there. And I believe we can even remove the LVM hack
for dm-crypt-over-LVM setups, too.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: not available
More information about the pkg-cryptsetup-devel