dmcrypt lockbox isn't cleared on reinstall

[Chris Sanders]

Testing dmcrypt with bluestore I've noticed that the lockbox fails to install when a drive was previously used as a dmcrypt bluestore device. I believe this is because the zap operation is not clearing the filesystem, the lockbox is re-created at the same location, and mk2fs stops for a user verification seeing an existing file system.

Running ceph-disk directly will produce the prompt at the mk2fs stage. The charmhelpers zap_disk function isn't sufficient, calling wipefs -a on the block device doesn't clear the necessary portion of the disk.

I've run across this working on other items and have a patch in progress. It appears you can solve this by calling wipefs on each partition before removing the partition. I'm opening this to raise awareness, but intend to fix this as a drive-by on another patch since I've touched this same part of the code in the other patch.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/533838

[OpenStack Infra (hudson-openstack)]

Change abandoned by Frode Nordahl (<email address hidden>) on branch: master
Review: https://review.openstack.org/533838
Reason: Marking this review as abandoned due to lack of activity in the past 6 months.
Feel free to restore it again if you want to pick up and continue the work.

[Chris Sanders (chris.sanders)]

I'm abandoning this and I'm not clear if it still applies with the latest charms change in OSD initialization.