Another step forward to having eqiad1 in production is to have the `nova` and `nova-api` databases in a production db server (m5-master).
Since the `nova` names are already in use by the old `main` deployment, we should pick another name, like `nova-eqiad1`/`nova-api-eqiad1` or something like that.
Steps:
[x] leave room for more nova connections
[x] review connectivity (should simply work after previous database migrations)
[x] m5-master DB GRANTS https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/455114/
[x] prepare puppet patch to refresh nova DB references for eqiad1 (server + db name)
[x] db backup-copy-import
[x] sync nova DB pass for main/eqiad1 (labs/private, etc)
[] merge puppet patch https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/454774/
[] Review m5 database backups so everything important is backed up
This task is for tracking such work.
For the record:
2 new DBs:
* `nova_eqiad1`
* `nova_api_eqiad1`
From:
* cloudcontrol1003.wikimedia.org / 208.80.154.23
* cloudcontrol1004.wikimedia.org / 208.80.154.132
I think the GRANTs would look like:
```
GRANT ALL PRIVILEGES ON nova_eqiad1.* TO 'nova'@'208.80.154.23' IDENTIFIED BY '<%= @nova_dbpass %>';
GRANT ALL PRIVILEGES ON nova_api_eqiad1.* TO 'nova'@'208.80.154.23' IDENTIFIED BY '<%= @nova_dbpass %>';
GRANT ALL PRIVILEGES ON nova_eqiad1.* TO 'nova'@'208.80.154.132' IDENTIFIED BY '<%= @nova_dbpass %>';
GRANT ALL PRIVILEGES ON nova_api_eqiad1.* TO 'nova'@'208.80.154.132' IDENTIFIED BY '<%= @nova_dbpass %>';
```
Password is `hieradata/eqiad/profile/openstack/eqiad1/nova.yaml` `profile::openstack::eqiad1::nova::db_pass` in the private repo.
(we can use the same as in `hieradata/eqiad/profile/openstack/main/nova.yaml` `profile::openstack::main::nova::db_pass`)