This is most likely a bit of a race between ceph-osd units being up and running, and the pool creation in ceph-mon occuring.
I think by default, ceph-mon does not wait until OSD units appear - but even that is not fail-safe as the osd's themselves might not be running.
This is most likely a bit of a race between ceph-osd units being up and running, and the pool creation in ceph-mon occuring.
I think by default, ceph-mon does not wait until OSD units appear - but even that is not fail-safe as the osd's themselves might not be running.