[integration tests] Nova live migrartion test TimeoutExpired: Timeout of 300 seconds expired waiting for instances to be ready
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
Georgy Dyuldin |
Bug Description
Test result:
https:/
Trace:
self = <mos_tests.
big_hypervisors = [<Hypervisor: 1>, <Hypervisor: 2>], block_migration = True
big_port_quota = None, with_volume = False
@pytest.
@pytest.
@pytest.
@pytest.
[(True, False), (False, False), (False, True)],
ids=['block LM w/o vol', 'true LM w/o vol', 'true LM w vol'],
@pytest.
def test_live_
self, big_hypervisors, block_migration, big_port_quota,
"""LM of maximum allowed amount of instances created with all available
1. Allow unlimited concurrent live migrations
2. Restart nova-api services on controllers and
3. Create maximum allowed number of instances on a single
4. Initiate serial block LM of previously created instances
5. Check that all live-migrated instances are hosted on target host
6. Send pings between pairs of VMs to check that network
7. Initiate concurrent block LM of previously created instances
8. Check that all live-migrated instances are hosted on target host
9. Send pings between pairs of VMs to check that network
10. Repeat pp.3-9 for every available flavor
"""
project_id = self.os_
image = self.os_
if with_volume:
for i in range(max_volumes):
vol = common.
zone = self.os_
flavors = sorted(
for flavor in flavors:
# Skip small flavors
if flavor.ram < 512:
if with_volume:
else:
> create_
mos_tests/
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <mos_tests.
zone = 'nova:node-
instances_count = 38, image_id = None, userdata = None
create_args = [{}, {}, {}, {}, {}, {}, ...]
def create_
boot_marker = 'INSTANCE BOOT COMPLETED'
if userdata is not None:
if create_args is not None:
assert len(create_args) == instances_count
else:
for i in range(instances
kwargs = create_args[i]
predicates = [lambda: self.os_
if userdata is None:
else:
> waiting_
E TimeoutExpired: Timeout of 300 seconds expired waiting for instances to be ready
mos_tests/
Changed in mos: | |
assignee: | nobody → MOS QA Team (mos-qa) |
Changed in mos: | |
milestone: | 10.0 → 9.0 |
Сannot reproduce it on my env, and see in http:// cz7776. bud.mirantis. net:8080/ jenkins/ job/9.0_ Neutron_ Nova/15/ testReport/ mos_tests. nova.live_ migration_ test/TestLiveMi gration/ test_live_ migration_ max_instances_ with_all_ flavors_ block_LM_ w_o_vol_ __838028_ _/ that booted instances went to ACTIVE status successfully, but a lot of "Instance unavailable yet: Error reading SSH protocol banner" and "Authentication failed." messages here for the 'waiting_ for="instances to be ready"' event. Also see that some of instances nevertheless got 'instances to be ready' status and were migrated. It looks like some env's resources/ configuration issues