diff options
Diffstat (limited to 'tests/cloud_tests/platforms')
| -rw-r--r-- | tests/cloud_tests/platforms/instances.py | 44 | ||||
| -rw-r--r-- | tests/cloud_tests/platforms/lxd/instance.py | 54 |
2 files changed, 75 insertions, 23 deletions
diff --git a/tests/cloud_tests/platforms/instances.py b/tests/cloud_tests/platforms/instances.py index 3bad021f..529e79cd 100644 --- a/tests/cloud_tests/platforms/instances.py +++ b/tests/cloud_tests/platforms/instances.py @@ -87,32 +87,39 @@ class Instance(TargetBase): self._ssh_client = None def _ssh_connect(self): - """Connect via SSH.""" + """Connect via SSH. + + Attempt to SSH to the client on the specific IP and port. If it + fails in some manner, then retry 2 more times for a total of 3 + attempts; sleeping a few seconds between attempts. + """ if self._ssh_client: return self._ssh_client if not self.ssh_ip or not self.ssh_port: - raise ValueError + raise ValueError("Cannot ssh_connect, ssh_ip=%s ssh_port=%s" % + (self.ssh_ip, self.ssh_port)) client = paramiko.SSHClient() client.set_missing_host_key_policy(paramiko.AutoAddPolicy()) private_key = paramiko.RSAKey.from_private_key_file(self.ssh_key_file) - retries = 30 + retries = 3 while retries: try: client.connect(username=self.ssh_username, hostname=self.ssh_ip, port=self.ssh_port, - pkey=private_key, banner_timeout=30) + pkey=private_key) self._ssh_client = client return client except (ConnectionRefusedError, AuthenticationException, BadHostKeyException, ConnectionResetError, SSHException, - OSError) as e: + OSError): retries -= 1 - time.sleep(10) + LOG.debug('Retrying ssh connection on connect failure') + time.sleep(3) - ssh_cmd = 'Failed ssh connection to %s@%s:%s after 300 seconds' % ( + ssh_cmd = 'Failed ssh connection to %s@%s:%s after 3 retries' % ( self.ssh_username, self.ssh_ip, self.ssh_port ) raise util.InTargetExecuteError(b'', b'', 1, ssh_cmd, 'ssh') @@ -128,18 +135,31 @@ class Instance(TargetBase): return ' '.join(l for l in test.strip().splitlines() if not l.lstrip().startswith('#')) - time = self.config['boot_timeout'] + boot_timeout = self.config['boot_timeout'] tests = [self.config['system_ready_script']] if wait_for_cloud_init: tests.append(self.config['cloud_init_ready_script']) formatted_tests = ' && '.join(clean_test(t) for t in tests) cmd = ('i=0; while [ $i -lt {time} ] && i=$(($i+1)); do {test} && ' - 'exit 0; sleep 1; done; exit 1').format(time=time, + 'exit 0; sleep 1; done; exit 1').format(time=boot_timeout, test=formatted_tests) - if self.execute(cmd, rcs=(0, 1))[-1] != 0: - raise OSError('timeout: after {}s system not started'.format(time)) - + end_time = time.time() + boot_timeout + while True: + try: + return_code = self.execute( + cmd, rcs=(0, 1), description='wait for instance start' + )[-1] + if return_code == 0: + break + except util.InTargetExecuteError: + LOG.warning("failed to connect via SSH") + + if time.time() < end_time: + time.sleep(3) + else: + raise util.PlatformError('ssh', 'after %ss instance is not ' + 'reachable' % boot_timeout) # vi: ts=4 expandtab diff --git a/tests/cloud_tests/platforms/lxd/instance.py b/tests/cloud_tests/platforms/lxd/instance.py index 0d957bca..83c97ab4 100644 --- a/tests/cloud_tests/platforms/lxd/instance.py +++ b/tests/cloud_tests/platforms/lxd/instance.py @@ -12,6 +12,8 @@ from tests.cloud_tests.util import PlatformError from ..instances import Instance +from pylxd import exceptions as pylxd_exc + class LXDInstance(Instance): """LXD container backed instance.""" @@ -30,6 +32,9 @@ class LXDInstance(Instance): @param config: image config @param features: supported feature flags """ + if not pylxd_container: + raise ValueError("Invalid value pylxd_container: %s" % + pylxd_container) self._pylxd_container = pylxd_container super(LXDInstance, self).__init__( platform, name, properties, config, features) @@ -40,9 +45,19 @@ class LXDInstance(Instance): @property def pylxd_container(self): """Property function.""" + if self._pylxd_container is None: + raise RuntimeError( + "%s: Attempted use of pylxd_container after deletion." % self) self._pylxd_container.sync() return self._pylxd_container + def __str__(self): + return ( + '%s(name=%s) status=%s' % + (self.__class__.__name__, self.name, + ("deleted" if self._pylxd_container is None else + self.pylxd_container.status))) + def _execute(self, command, stdin=None, env=None): if env is None: env = {} @@ -152,9 +167,8 @@ class LXDInstance(Instance): return fp.read() try: - stdout, stderr = subp( - ['lxc', 'console', '--show-log', self.name], decode=False) - return stdout + return subp(['lxc', 'console', '--show-log', self.name], + decode=False)[0] except ProcessExecutionError as e: raise PlatformError( "console log", @@ -166,10 +180,27 @@ class LXDInstance(Instance): self.shutdown(wait=wait) self.start(wait=wait) - def shutdown(self, wait=True): + def shutdown(self, wait=True, retry=1): """Shutdown instance.""" - if self.pylxd_container.status != 'Stopped': + if self.pylxd_container.status == 'Stopped': + return + + try: + LOG.debug("%s: shutting down (wait=%s)", self, wait) self.pylxd_container.stop(wait=wait) + except (pylxd_exc.LXDAPIException, pylxd_exc.NotFound) as e: + # An exception happens here sometimes (LP: #1783198) + # LOG it, and try again. + LOG.warning( + ("%s: shutdown(retry=%d) caught %s in shutdown " + "(response=%s): %s"), + self, retry, e.__class__.__name__, e.response, e) + if isinstance(e, pylxd_exc.NotFound): + LOG.debug("container_exists(%s) == %s", + self.name, self.platform.container_exists(self.name)) + if retry == 0: + raise e + return self.shutdown(wait=wait, retry=retry - 1) def start(self, wait=True, wait_for_cloud_init=False): """Start instance.""" @@ -190,12 +221,14 @@ class LXDInstance(Instance): def destroy(self): """Clean up instance.""" + LOG.debug("%s: deleting container.", self) self.unfreeze() self.shutdown() self.pylxd_container.delete(wait=True) + self._pylxd_container = None + if self.platform.container_exists(self.name): - raise OSError('container {} was not properly removed' - .format(self.name)) + raise OSError('%s: container was not properly removed' % self) if self._console_log_file and os.path.exists(self._console_log_file): os.unlink(self._console_log_file) shutil.rmtree(self.tmpd) @@ -209,16 +242,15 @@ def _has_proper_console_support(): if 'console' not in info.get('api_extensions', []): reason = "LXD server does not support console api extension" else: - dver = info.get('environment', {}).get('driver_version', "") + dver = str(info.get('environment', {}).get('driver_version', "")) if dver.startswith("2.") or dver.startswith("1."): reason = "LXD Driver version not 3.x+ (%s)" % dver else: try: - stdout, stderr = subp(['lxc', 'console', '--help'], - decode=False) + stdout = subp(['lxc', 'console', '--help'], decode=False)[0] if not (b'console' in stdout and b'log' in stdout): reason = "no '--log' in lxc console --help" - except ProcessExecutionError as e: + except ProcessExecutionError: reason = "no 'console' command in lxc client" if reason: |