[Bug 1779685] Re: [18.04.1] Backport support for Intel VROC arrays in mdadm

[Steve Langasek]

reviewing the diff, my biggest concern is that there appear to be new
udev rules upstream.  Are we going to be shipping these new upstream
udev rules in the mdadm package?  Have you examined the impact on other
block layers, in particular lvm where I believe we are shipping non-
upstream rules?

** Changed in: mdadm (Ubuntu Bionic)
       Status: In Progress => Incomplete

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to mdadm in Ubuntu.
https://bugs.launchpad.net/bugs/1779685

Title:
  [18.04.1] Backport support for Intel VROC arrays in mdadm

Status in mdadm package in Ubuntu:
  Fix Released
Status in mdadm source package in Bionic:
  Incomplete
Status in mdadm source package in Cosmic:
  Fix Released

Bug description:
  [Impact]

   * mdadm as shipped in 18.04.0 does not support Intel VROC arrays, as
  to be shipped over Bionic LTS lifecycle in multiple
  laptop/desktop/server SKUs by multiple OEMs.

   * this is a request to backport mdadm master, with support of Intel
  VROC arrays, and subsequent compiler fixes, and reshaping fixes, to
  fully support VROC arrays (including reshaping).

  [Test Cases]

   * Ensure existing Intel IMSM raids continue to assemble, boot, shutdown cleanly, bootable when degraded.
   * Ensure existing linux raids continue to assemble, boot, shutdown cleanly, bootable when degraded.
     - with 0.90 metadata; 1.0 metadata; 1.1 metadata; 1.2 metadata
   * [Optional] ensure that DDF raids continue to assemble, boot, shutdown cleanly, bootable when degraded.
   * Validate installers onto Intel IMSM raids
   * Validate installers onto Linux RAID
   * Validate bare-metal public clouds that use mdadm

  [Regression Potential]

   * Worst case scenario is failing to start assembled array in rw mode,
  on boot. This would result in dropping into emergency shell/mode in
  either initramfs, or emergency.target requiring manual intervention by
  a sysadmin to recover and start synced raid array. However, the test
  cases above should cover this regression potential adequately.

   * Other changes, to resizing/reshaping may result in failure to
  resize/reshape, but these are rare maintenance operations for which
  one should have a backup - or able to recover from such an operation.

  [Other Info]

   * xnox offline 7th July - 16th July

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/mdadm/+bug/1779685/+subscriptions