[9.x acceptance] Time limit exceeded while waiting for instance in "active" state

Detailed bug description:

The issue was found by https://product-ci.infra.mirantis.net/job/9.x.acceptance.ubuntu.ha_scale_group_5/40/testReport/(root)/add_delete_compute_cinder_ceph/

Steps to reproduce:
        1. Deploy cluster 3 controllers, 2 computes + ceph + cinder,
           Neutron VXLAN, cinder for volumes, ceph for images.
        2. Verify networks
        3. Run OSTF

OSTF test "Create volume and boot instance from it" fails with
"Time limit exceeded while waiting for instance becoming 'ACTIVE' to finish"

Some part of logs http://paste.openstack.org/show/596904/

additional info:
The failed test had passed after env reverting