x
/
tobiko
Code Issues Proposed changes

add a reset_controller_main_vip disruptive test 

Change-Id: If3b741e1372dcd43185663d78cd5f08c05e965f6
This commit is contained in:
pinikomarov 2020-04-28 09:10:05 +03:00
parent 45202f038e
commit 5dc4dfb214
1 changed files with 55 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
tobiko/tests/faults/ha/cloud_disruptions.py
View File

@ -1,25 +1,59 @@
from __future__ import absolute_import
import re
import tobiko
from tobiko.shell import sh
from tobiko.openstack import topology
from tobiko.tripleo import topology as tripleo_topology
from tobiko.openstack import keystone
from tobiko.tripleo import pacemaker
from oslo_log import log
LOG = log.getLogger(__name__)
hard_reset_method = 'sudo chmod o+w /proc/sysrq-trigger;' \
'sudo echo b > /proc/sysrq-trigger'
soft_reset_method = 'sudo reboot'
def get_node(node_name):
return [node for node in topology.list_openstack_nodes() if
node.name == node_name][0]
def reset_node(node_name, hard_reset=False):
# reboot all controllers and wait for ssh Up on them
# hard reset is simultaneous while soft is sequential
if hard_reset:
reset_method = hard_reset_method
else:
reset_method = soft_reset_method
# using ssh_client.connect we use a fire and forget reboot method
node = get_node(node_name)
node.ssh_client.connect().exec_command(reset_method)
LOG.info('reboot exec: {} on server: {}'.format(reset_method,
node.name))
tobiko.cleanup_fixture(node.ssh_client)
node_checked = sh.execute("hostname",
ssh_client=node.ssh_client,
expect_exit_status=None).stdout
LOG.info('{} is up '.format(node_checked))
def reset_all_controller_nodes(hard_reset=False):
# reboot all controllers and wait for ssh Up on them
# hard reset is simultaneous while soft is sequential
if hard_reset:
reset_method = 'sudo chmod o+w /proc/sysrq-trigger;' \
'sudo echo b > /proc/sysrq-trigger'
reset_method = hard_reset_method
else:
reset_method = 'sudo reboot'
reset_method = soft_reset_method
controlplane_groups = ['controller', 'messaging', 'database', 'networker']
actual_controlplane_groups = tripleo_topology.actual_node_groups(
controlplane_groups)
@ -38,15 +72,30 @@ def reset_all_controller_nodes(hard_reset=False):
LOG.info('{} is up '.format(controller_checked))
def reset_controller_main_vip(hard_reset=True):
# reset the controller holding the main vip (os_auth_url)
# ip resource (managed via pacemaker)
# find main vip by getting it from
credentials = keystone.default_keystone_credentials()
auth_url = credentials.auth_url
auth_url_ip = re.findall(r'[0-9]+(?:\.[0-9]+){3}', auth_url)[0]
# find the node holding that resource via :
main_vim_controller = pacemaker.get_overcloud_nodes_running_pcs_resource(
resource=f"ip-{auth_url_ip}")[0]
# get that node's ssh_client and reset it
reset_node(main_vim_controller, hard_reset=hard_reset)
def reset_all_compute_nodes(hard_reset=False):
# reboot all computes and wait for ssh Up on them
# hard reset is simultaneous while soft is sequential
if hard_reset:
reset_method = 'sudo chmod o+w /proc/sysrq-trigger;' \
'sudo echo b > /proc/sysrq-trigger'
reset_method = hard_reset_method
else:
reset_method = 'sudo reboot'
reset_method = soft_reset_method
for compute in topology.list_openstack_nodes(group='compute'):
# using ssh_client.connect we use a fire and forget reboot method
compute.ssh_client.connect().exec_command(reset_method)