System tests for k8s deploy have incorrect check for etcd nodes number

Steps to reproduce:

1. Create a fuel-devops YAML template for 5 nodes (default.yaml have only 3 nodes)
2. Use it for run fuel-ccp-tests:
export CONF_PATH="default_5_nodes.yaml"
...
py.test fuel_ccp_tests/tests/system/test_ccp_install_k8s.py::TestFuelCCPInstaller::test_k8s_installed_default

Expected result:
- test passed

Actual result:
- deploy passed, but check for etcd nodes failed:

    def check_etcd_health(self, underlay):
        node_names = underlay.node_names()
        cmd = "etcdctl cluster-health | grep -c 'got healthy result'"

        etcd_nodes = underlay.sudo_check_call(
            cmd, node_name=node_names[0])['stdout'][0]
> assert int(etcd_nodes) == len(node_names),\
            "Number of etcd nodes is {0}," \
            " should be {1}".format(int(etcd_nodes), len(node_names))
E AssertionError: Number of etcd nodes is 3, should be 5
E assert 3 == 5
E + where 3 = int('3\n')
E + and 5 = len(['node-1', 'node-2', 'node-3', 'node-4', 'node-5'])

------------
Possible solution:
It is impossible to set necessary amount of 'etcd' nodes before deploy (see the bug https://bugs.launchpad.net/fuel-ccp/+bug/1619735), so the necessary information can be taken from the inventory.cfg, that was used for deploy, on the kube master node:

$ /home/vagrant/workspace/inventory/inventory.cfg