nsxv + ceph deploy failed

Fuel 9.1, nsxv 3.0.1 #111

Steps to reproduce:
  1. Install plugin to the master node.
  2. Create cluster with vcenter and nsxv, setup networks and plugin settings.
  3. Add nodes with following roles:
       * 3 controllers
       * 3 ceph
       * 1 compute-vmware
       * 1 cinder-vmware
  4. Deploy cluster.

Expected result:
  Deploy passed

Actual result:
  Provisioning freezes at 100% for a long time, then deploy continues and then fails with errors on ceph-nodes.

  Part of puppet logs:
Puppet (debug): Executing 'ceph-disk activate-all'
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice): mount: wrong fs type, bad option, bad superblock on /dev/vdc3,
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice): missing codepage or helper program, or other error
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice): In some cases useful info is found in syslog - try
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice): dmesg | tail or so
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice):
/Stage[main]/Ceph::Osds/Exec[ceph-disk activate-all]/returns (notice): ceph-disk: Mounting filesystem failed: Command '['/bin/mount', '-t', 'xfs', '-o', 'rw,relatime,ize=256k,delaylog,allocsize=4M', '--', '/dev/disk/by-parttypeuuid/4fbd7e29-...819', '/var/lib/ceph/tmp/mnt.7N0OAD']' returned non-zero32

  Part of dmesg:
XFS (vdb3): unknown mount option [delaylog]