Hi,
We are currently running an Ocata "all-in-one" RDO instance within our hyperconverged win2016 failover cluster (S2D + HyperV). We recently added a new hyperv but we are having some issues...
We had previously used an old hyperv node back when we started our project but since then, the old node has been decomissioned and deleted from openstack through horizon.
That new hyperv node is using the same FQDN which appears to be causing some issues...
We are seeing the following errors within the "nova-compute.log" logfile:
2018-06-28 10:31:14.536 4796 WARNING nova.compute.manager [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] Instance e016c381-aab7-4668-8cea-3895d7d574b3 appears to not be owned by this host, but by HYPER04. Startup processing is being skipped.
2018-06-28 10:31:14.614 4796 ERROR nova.compute.manager [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] No compute node record for host hyper04
2018-06-28 10:31:14.628 4796 WARNING nova.compute.monitors [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] Excluding nova.compute.monitors.cpu monitor virt_driver. Not in the list of enabled monitors (CONF.compute_monitors).
2018-06-28 10:31:14.801 4796 WARNING nova.compute.resource_tracker [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] No compute node record for hyper04:hyper04
2018-06-28 10:31:16.770 4796 WARNING nova.scheduler.client.report [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] Unable to refresh my resource provider record
2018-06-28 10:31:16.959 4796 WARNING nova.scheduler.client.report [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] Unable to submit allocation for instance e016c381-aab7-4668-8cea-3895d7d574b3 (400 400 Bad Request
The server could not comply with the request since it is either malformed or otherwise incorrect.
Allocation for resource provider '6417dc89-882b-4cca-96d1-d20eb7f7757d' that does not exist. )
2018-06-28 10:31:17.036 4796 WARNING nova.compute.resource_tracker [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] [instance: e016c381-aab7-4668-8cea-3895d7d574b3] Instance not resizing, skipping migration.
2018-06-28 10:31:17.209 4796 WARNING nova.scheduler.client.report [req-39d711db-5e98-4e06-a637-4984311f2736 - - - - -] Unable to refresh my resource provider record
We put the newly added host in maintenance, stopped openstack services (nova, ceilometer and neutron) and then tried to remove the "hyper04" entry from the hostmappings table (novaapi's mysql DB) as well as removed the entry from the compute_nodes table (nova's mysql DB).
We then tried to re-map the host to cells:
# nova-manage cell_v2 discover_hosts --verbose
Found 2 cell mappings.
Skipping cell0 since it does not contain hosts.
Getting compute nodes from cell 'default': 64d7bbe9-b9e9-470f-8f96-8403dc3321f3
Found 5 computes in cell: 64d7bbe9-b9e9-470f-8f96-8403dc3321f3
Checking host mapping for compute host 'openstack-controller-blablabla.blablabla.com': 78fde1b7-b5d6-40d3-bbae-46e241140163
Checking host mapping for compute host 'hyper05': bf14ea3a-84a4-4f1e-8862-bbcf45950390
Checking host mapping for compute host 'hyper06': ec2a8ec3-33f7-4f35-bd84-617e32623514
Checking host mapping for compute host 'hyper07': fcd0409b-7560-42ee-9c19-027e92929822
Checking host mapping for compute host 'hyper04': 6417dc89-882b-4cca-96d1-d20eb7f7757d
Creating host mapping for compute host 'hyper04': 6417dc89-882b-4cca-96d1-d20eb7f7757d
However, the same errors appear in the nova's logfile.
Are we still missing something ? Is there any other entry that should be removed from the mysql DBs ?
