I've hacked in a udevadm settle call into the ceph-osd vaultlocker code:
juju upgrade-charm --switch cs:~james-page/ceph-osd-bug1783308 ceph-osd
This will prove if @dmitriis hypothesis about udev racing is correct; we can then investigate a more complete solution.
I've hacked in a udevadm settle call into the ceph-osd vaultlocker code:
juju upgrade-charm --switch cs:~james- page/ceph- osd-bug1783308 ceph-osd
This will prove if @dmitriis hypothesis about udev racing is correct; we can then investigate a more complete solution.