def azure_post_create_steps(self, email, backend_id, machine_id, monitoring, command,
key_id, username, password, public_key):
from mist.io.methods import ssh_command, connect_provider, enable_monitoring
from mist.io.methods import notify_user, notify_admin
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=120, max_retries=5)
try:
#login with user, password. Deploy the public key, enable sudo access for
#username, disable password authentication and reload ssh.
#After this is done, call post_deploy_steps if deploy script or monitoring
#is provided
ssh=paramiko.SSHClient()
ssh.load_system_host_keys()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect(host, username=username, password=password)
ssh.exec_command('mkdir -p ~/.ssh && echo "%s" >> ~/.ssh/authorized_keys && chmod -R 700 ~/.ssh/' % public_key)
chan = ssh.invoke_shell()
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command('sudo su -c \'echo "%s ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers\' ' % username)
chan.send('%s\n' % password)
chan = ssh.invoke_shell()
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command('sudo su -c \'echo "%s ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers.d/waagent\' ' % username)
chan.send('%s\n' % password)
cmd = 'sudo su -c \'sed -i "s|[#]*PasswordAuthentication yes|PasswordAuthentication no|g" /etc/ssh/sshd_config && /etc/init.d/ssh reload; service ssh reload\' '
ssh.exec_command(cmd)
ssh.close()
if command or monitoring:
post_deploy_steps.delay(email, backend_id, machine_id,
monitoring, command, key_id)
except Exception as exc:
raise self.retry(exc=exc, countdown=60, max_retries=10)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def hpcloud_post_create_steps(self, email, backend_id, machine_id, monitoring,
command, key_id, username, password, public_key,
script_id='', script_params='', job_id=None,
hostname='', plugins=None,
post_script_id='', post_script_params=''):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
else:
try:
available_networks = conn.ex_list_networks()
external_networks = [net for net in available_networks if net.router_external]
if external_networks:
ext_net_id = external_networks[0].id
else:
ext_net_id = ""
ports = conn.ex_list_ports()
port = [port for port in ports if port.get("device_id", "") == node.id][0]
ip = conn.ex_create_floating_ip(ext_net_id, port['id'])
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except:
raise self.retry(exc=Exception(), max_retries=20)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def rackspace_first_gen_post_create_steps(
self, email, cloud_id, machine_id, monitoring, command, key_id,
password, public_key, username='******', script_id='', script_params='',
job_id=None, hostname='', plugins=None, post_script_id='',
post_script_params='', cronjob={}
):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.clouds[cloud_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), max_retries=20)
try:
# login with user, password and deploy the ssh public key.
# Disable password authentication and reload ssh.
# After this is done, call post_deploy_steps
# if deploy script or monitoring
# is provided
ssh=paramiko.SSHClient()
ssh.load_system_host_keys()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect(host, username=username, password=password, timeout=None, allow_agent=False, look_for_keys=False)
ssh.exec_command('mkdir -p ~/.ssh && echo "%s" >> ~/.ssh/authorized_keys && chmod -R 700 ~/.ssh/' % public_key)
cmd = 'sudo su -c \'sed -i "s|[#]*PasswordAuthentication yes|PasswordAuthentication no|g" /etc/ssh/sshd_config && /etc/init.d/ssh reload; service ssh reload\' '
ssh.exec_command(cmd)
ssh.close()
post_deploy_steps.delay(
email, cloud_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params, cronjob=cronjob
)
except Exception as exc:
raise self.retry(exc=exc, countdown=10, max_retries=15)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def rackspace_first_gen_post_create_steps(
self, email, backend_id, machine_id, monitoring, command, key_id,
password, public_key, username='******', script_id='', script_params='',
job_id=None, hostname='', plugins=None, post_script_id='',
post_script_params=''
):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), max_retries=20)
try:
# login with user, password and deploy the ssh public key. Disable password authentication and reload ssh.
# After this is done, call post_deploy_steps if deploy script or monitoring
# is provided
ssh=paramiko.SSHClient()
ssh.load_system_host_keys()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect(host, username=username, password=password, timeout=None, allow_agent=False, look_for_keys=False)
ssh.exec_command('mkdir -p ~/.ssh && echo "%s" >> ~/.ssh/authorized_keys && chmod -R 700 ~/.ssh/' % public_key)
cmd = 'sudo su -c \'sed -i "s|[#]*PasswordAuthentication yes|PasswordAuthentication no|g" /etc/ssh/sshd_config && /etc/init.d/ssh reload; service ssh reload\' '
ssh.exec_command(cmd)
ssh.close()
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except Exception as exc:
raise self.retry(exc=exc, countdown=10, max_retries=15)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def run_deploy_script(self, email, backend_id, machine_id, command,
key_id=None, username=None, password=None, port=22):
from mist.io.methods import ssh_command, connect_provider
from mist.io.methods import notify_user, notify_admin
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=60, max_retries=5)
try:
from mist.io.shell import Shell
shell = Shell(host)
key_id, ssh_user = shell.autoconfigure(user, backend_id, node.id,
key_id, username, password, port)
start_time = time()
retval, output = shell.command(command)
execution_time = time() - start_time
shell.disconnect()
msg = """
Command: %s
Return value: %s
Duration: %s seconds
Output:
%s""" % (command, retval, execution_time, output)
if retval:
notify_user(user, "Deployment script failed for machine %s (%s)" % (node.name, node.id), msg)
amqp_log("Deployment script failed for user %s machine %s (%s): %s" % (user, node.name, node.id, msg))
else:
notify_user(user, "Deployment script succeeded for machine %s (%s)" % (node.name, node.id), msg)
amqp_log("Deployment script succeeded for user %s machine %s (%s): %s" % (user, node.name, node.id, msg))
except ServiceUnavailableError as exc:
raise self.retry(exc=exc, countdown=60, max_retries=5)
except Exception as exc:
if str(exc).startswith('Retry'):
return
amqp_log("Deployment script failed for machine %s in backend %s by user %s after 5 retries: %s" % (node.id, backend_id, email, repr(exc)))
notify_user(user, "Deployment script failed for machine %s after 5 retries" % node.id)
notify_admin("Deployment script failed for machine %s in backend %s by user %s after 5 retries" % (node.id, backend_id, email), repr(exc))
def post_deploy_steps(self, email, backend_id, machine_id, monitoring, command,
key_id=None, username=None, password=None, port=22):
from mist.io.methods import ssh_command, connect_provider, enable_monitoring
from mist.io.methods import notify_user, notify_admin
if multi_user:
from mist.core.methods import enable_monitoring
else:
from mist.io.methods import enable_monitoring
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=120, max_retries=5)
try:
from mist.io.shell import Shell
shell = Shell(host)
# connect with ssh even if no command, to create association
# to be able to enable monitoring
key_id, ssh_user = shell.autoconfigure(
user, backend_id, node.id, key_id, username, password, port
)
if command:
start_time = time()
retval, output = shell.command(command)
execution_time = time() - start_time
output = output.decode('utf-8','ignore')
msg = ("Command: %s\n"
"Return value: %s\n"
"Duration: %d seconds\n"
"Output:%s\n") % (command, retval,
execution_time, output)
msg = msg.encode('utf-8', 'ignore')
msg_title = "Deployment script %s for machine %s (%s)" % (
'failed' if retval else 'succeeded',
node.name, node.id
)
notify_user(user, msg_title, msg)
shell.disconnect()
if monitoring:
try:
enable_monitoring(user, backend_id, node.id,
name=node.name, dns_name=node.extra.get('dns_name',''),
public_ips=ips, no_ssh=False, dry=False,
)
except Exception as e:
print repr(e)
notify_user(user, "Enable monitoring failed for machine %s (%s)" % (node.name, node.id), repr(e))
notify_admin('Enable monitoring on creation failed for user %s machine %s: %r' % (email, node.name, e))
except (ServiceUnavailableError, SSHException) as exc:
raise self.retry(exc=exc, countdown=60, max_retries=5)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
amqp_log("Deployment script failed for machine %s in backend %s by user %s after 5 retries: %s" % (node.id, backend_id, email, repr(exc)))
notify_user(user, "Deployment script failed for machine %s after 5 retries" % node.id)
notify_admin("Deployment script failed for machine %s in backend %s by user %s after 5 retries" % (node.id, backend_id, email), repr(exc))
def post_deploy_steps(self, email, backend_id, machine_id, monitoring, command,
key_id=None, username=None, password=None, port=22,
script_id='', script_params='', job_id=None,
hostname='', plugins=None,
post_script_id='', post_script_params=''):
from mist.io.methods import connect_provider, probe_ssh_only
from mist.io.methods import notify_user, notify_admin
from mist.io.methods import create_dns_a_record
if multi_user:
from mist.core.methods import enable_monitoring
from mist.core.tasks import run_script
from mist.core.helpers import log_event
else:
from mist.io.methods import enable_monitoring
log_event = lambda *args, **kwargs: None
job_id = job_id or uuid.uuid4().hex
user = user_from_email(email)
tmp_log = lambda msg, *args: log.error('Post deploy: %s' % msg, *args)
tmp_log('Entering post deploy steps for %s %s %s',
user.email, backend_id, machine_id)
try:
# find the node we're looking for and get its hostname
node = None
try:
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes() # TODO: use cache
for n in nodes:
if n.id == machine_id:
node = n
break
tmp_log('run list_machines')
except:
raise self.retry(exc=Exception(), countdown=10, max_retries=10)
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
tmp_log('ip not found, retrying')
raise self.retry(exc=Exception(), countdown=60, max_retries=20)
try:
from mist.io.shell import Shell
shell = Shell(host)
# connect with ssh even if no command, to create association
# to be able to enable monitoring
tmp_log('attempting to connect to shell')
key_id, ssh_user = shell.autoconfigure(
user, backend_id, node.id, key_id, username, password, port
)
tmp_log('connected to shell')
result = probe_ssh_only(user, backend_id, machine_id, host=None,
key_id=key_id, ssh_user=ssh_user, shell=shell)
log_dict = {
'email': email,
'event_type': 'job',
'backend_id': backend_id,
'machine_id': machine_id,
'job_id': job_id,
'host': host,
'key_id': key_id,
'ssh_user': ssh_user,
}
log_event(action='probe', result=result, **log_dict)
backend = user.backends[backend_id]
msg = "Backend:\n Name: %s\n Id: %s\n" % (backend.title,
backend_id)
msg += "Machine:\n Name: %s\n Id: %s\n" % (node.name,
node.id)
if hostname:
try:
record = create_dns_a_record(user, hostname, host)
hostname = '.'.join((record.name, record.zone.domain))
log_event(action='create_dns_a_record', hostname=hostname,
**log_dict)
except Exception as exc:
log_event(action='create_dns_a_record', error=str(exc),
**log_dict)
error = False
if script_id and multi_user:
tmp_log('will run script_id %s', script_id)
ret = run_script.run(
user.email, script_id, backend_id, machine_id,
params=script_params, host=host, job_id=job_id
)
error = ret['error']
tmp_log('executed script_id %s', script_id)
elif command:
tmp_log('will run command %s', command)
log_event(action='deployment_script_started', command=command, **log_dict)
start_time = time()
retval, output = shell.command(command)
tmp_log('executed command %s', command)
execution_time = time() - start_time
output = output.decode('utf-8','ignore')
title = "Deployment script %s" % ('failed' if retval
else 'succeeded')
error = retval > 0
notify_user(user, title,
backend_id=backend_id,
machine_id=machine_id,
machine_name=node.name,
command=command,
output=output,
duration=execution_time,
retval=retval,
error=retval > 0)
log_event(action='deployment_script_finished',
error=retval > 0,
return_value=retval,
command=command,
stdout=output,
**log_dict)
shell.disconnect()
if monitoring:
try:
enable_monitoring(user, backend_id, node.id,
name=node.name, dns_name=node.extra.get('dns_name',''),
public_ips=ips, no_ssh=False, dry=False, job_id=job_id,
plugins=plugins, deploy_async=False,
)
except Exception as e:
print repr(e)
error = True
notify_user(user, "Enable monitoring failed for machine %s" % machine_id, repr(e))
notify_admin('Enable monitoring on creation failed for user %s machine %s: %r' % (email, machine_id, e))
log_event(action='enable_monitoring_failed', error=repr(e),
**log_dict)
if post_script_id and multi_user:
tmp_log('will run post_script_id %s', post_script_id)
ret = run_script.run(
user.email, post_script_id, backend_id, machine_id,
params=post_script_params, host=host, job_id=job_id,
action_prefix='post_',
)
error = ret['error']
tmp_log('executed post_script_id %s', script_id)
log_event(action='post_deploy_finished', error=error, **log_dict)
except (ServiceUnavailableError, SSHException) as exc:
tmp_log(repr(exc))
raise self.retry(exc=exc, countdown=60, max_retries=15)
except Exception as exc:
tmp_log(repr(exc))
if str(exc).startswith('Retry'):
raise
notify_user(user, "Deployment script failed for machine %s" % machine_id)
notify_admin("Deployment script failed for machine %s in backend %s by user %s" % (machine_id, backend_id, email), repr(exc))
log_event(
email=email,
event_type='job',
action='post_deploy_finished',
backend_id=backend_id,
machine_id=machine_id,
enable_monitoring=bool(monitoring),
command=command,
error="Couldn't connect to run post deploy steps.",
job_id=job_id
)
def azure_post_create_steps(self, email, backend_id, machine_id, monitoring,
command, key_id, username, password, public_key,
script_id='', script_params='', job_id=None,
hostname='', plugins=None,
post_script_id='', post_script_params=''):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), max_retries=20)
try:
# login with user, password. Deploy the public key, enable sudo access for
# username, disable password authentication and reload ssh.
# After this is done, call post_deploy_steps if deploy script or monitoring
# is provided
ssh = paramiko.SSHClient()
ssh.load_system_host_keys()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect(host, username=username, password=password, timeout=None, allow_agent=False, look_for_keys=False)
ssh.exec_command('mkdir -p ~/.ssh && echo "%s" >> ~/.ssh/authorized_keys && chmod -R 700 ~/.ssh/' % public_key)
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command('sudo su -c \'echo "%s ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers\' ' % username)
chan.send('%s\n' % password)
check_sudo_command = 'sudo su -c \'whoami\''
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command(check_sudo_command)
output = chan.recv(1024)
if not output.startswith('root'):
raise
cmd = 'sudo su -c \'sed -i "s|[#]*PasswordAuthentication yes|PasswordAuthentication no|g" /etc/ssh/sshd_config && /etc/init.d/ssh reload; service ssh reload\' '
ssh.exec_command(cmd)
ssh.close()
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except Exception as exc:
raise self.retry(exc=exc, countdown=10, max_retries=15)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def openstack_post_create_steps(self, email, backend_id, machine_id, monitoring,
command, key_id, username, password, public_key,
script_id='', script_params='', job_id=None,
hostname='', plugins=None,
post_script_id='', post_script_params='', networks=[]):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
else:
try:
created_floating_ips = []
for network in networks['public']:
created_floating_ips += [floating_ip for floating_ip in network['floating_ips']]
# From the already created floating ips try to find one that is not associated to a node
unassociated_floating_ip = None
for ip in created_floating_ips:
if not ip['node_id']:
unassociated_floating_ip = ip
break
# Find the ports which are associated to the machine (e.g. the ports of the private ips)
# and use one to associate a floating ip
ports = conn.ex_list_ports()
machine_port_id = None
for port in ports:
if port.get('device_id') == node.id:
machine_port_id = port.get('id')
break
if unassociated_floating_ip:
log.info("Associating floating ip with machine: %s" % node.id)
ip = conn.ex_associate_floating_ip_to_node(unassociated_floating_ip['id'], machine_port_id)
else:
# Find the external network
log.info("Create and associating floating ip with machine: %s" % node.id)
ext_net_id = networks['public'][0]['id']
ip = conn.ex_create_floating_ip(ext_net_id, machine_port_id)
post_deploy_steps.delay(
email, backend_id, machine_id, monitoring, command, key_id,
script_id=script_id, script_params=script_params,
job_id=job_id, hostname=hostname, plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except:
raise self.retry(exc=Exception(), max_retries=20)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def post_deploy_steps(self,
email,
backend_id,
machine_id,
monitoring,
command,
key_id=None,
username=None,
password=None,
port=22):
from mist.io.methods import ssh_command, connect_provider, enable_monitoring
from mist.io.methods import notify_user, notify_admin
if multi_user:
from mist.core.methods import enable_monitoring
from mist.core.helpers import log_event
else:
from mist.io.methods import enable_monitoring
log_event = lambda *args, **kwargs: None
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=120, max_retries=5)
try:
from mist.io.shell import Shell
shell = Shell(host)
# connect with ssh even if no command, to create association
# to be able to enable monitoring
key_id, ssh_user = shell.autoconfigure(user, backend_id, node.id,
key_id, username, password,
port)
backend = user.backends[backend_id]
msg = "Backend:\n Name: %s\n Id: %s\n" % (backend.title,
backend_id)
msg += "Machine:\n Name: %s\n Id: %s\n" % (node.name, node.id)
if command:
log_dict = {
'email': email,
'event_type': 'job',
'backend_id': backend_id,
'machine_id': machine_id,
'job_id': uuid.uuid4().hex,
'command': command,
'host': host,
'key_id': key_id,
'ssh_user': ssh_user,
}
log_event(action='deployment_script_started', **log_dict)
start_time = time()
retval, output = shell.command(command)
execution_time = time() - start_time
output = output.decode('utf-8', 'ignore')
title = "Deployment script %s" % ('failed'
if retval else 'succeeded')
notify_user(user,
title,
backend_id=backend_id,
machine_id=machine_id,
machine_name=node.name,
command=command,
output=output,
duration=execution_time,
retval=retval,
error=retval > 0)
log_event(action='deployment_script_finished',
error=retval > 0,
return_value=retval,
stdout=output,
**log_dict)
shell.disconnect()
if monitoring:
try:
enable_monitoring(
user,
backend_id,
node.id,
name=node.name,
dns_name=node.extra.get('dns_name', ''),
public_ips=ips,
no_ssh=False,
dry=False,
)
except Exception as e:
print repr(e)
notify_user(
user, "Enable monitoring failed for machine %s (%s)" %
(node.name, node.id), repr(e))
notify_admin(
'Enable monitoring on creation failed for user %s machine %s: %r'
% (email, node.name, e))
except (ServiceUnavailableError, SSHException) as exc:
raise self.retry(exc=exc, countdown=60, max_retries=5)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
notify_user(
user, "Deployment script failed for machine %s after 5 retries" %
node.id)
notify_admin(
"Deployment script failed for machine %s in backend %s by user %s after 5 retries"
% (node.id, backend_id, email), repr(exc))
log_event(
email=email,
event_type='job',
action='deployment_script_failed',
backend_id=backend_id,
machine_id=machine_id,
enable_monitoring=bool(monitoring),
command=command,
error="Couldn't connect to run post deploy steps (5 attempts).",
)
def azure_post_create_steps(self, email, backend_id, machine_id, monitoring,
command, key_id, username, password, public_key):
from mist.io.methods import ssh_command, connect_provider, enable_monitoring
from mist.io.methods import notify_user, notify_admin
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), max_retries=20)
try:
# login with user, password. Deploy the public key, enable sudo access for
# username, disable password authentication and reload ssh.
# After this is done, call post_deploy_steps if deploy script or monitoring
# is provided
ssh = paramiko.SSHClient()
ssh.load_system_host_keys()
ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
ssh.connect(host,
username=username,
password=password,
timeout=None,
allow_agent=False,
look_for_keys=False)
ssh.exec_command(
'mkdir -p ~/.ssh && echo "%s" >> ~/.ssh/authorized_keys && chmod -R 700 ~/.ssh/'
% public_key)
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command(
'sudo su -c \'echo "%s ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers\' '
% username)
chan.send('%s\n' % password)
check_sudo_command = 'sudo su -c \'whoami\''
chan = ssh.get_transport().open_session()
chan.get_pty()
chan.exec_command(check_sudo_command)
output = chan.recv(1024)
if not output.startswith('root'):
raise
cmd = 'sudo su -c \'sed -i "s|[#]*PasswordAuthentication yes|PasswordAuthentication no|g" /etc/ssh/sshd_config && /etc/init.d/ssh reload; service ssh reload\' '
ssh.exec_command(cmd)
ssh.close()
if command or monitoring:
post_deploy_steps.delay(email, backend_id, machine_id,
monitoring, command, key_id)
except Exception as exc:
raise self.retry(exc=exc, countdown=10, max_retries=15)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def run_deploy_script(self,
email,
backend_id,
machine_id,
command,
key_id=None,
username=None,
password=None,
port=22):
from mist.io.methods import ssh_command, connect_provider
from mist.io.methods import notify_user, notify_admin
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=60, max_retries=5)
try:
from mist.io.shell import Shell
shell = Shell(host)
key_id, ssh_user = shell.autoconfigure(user, backend_id, node.id,
key_id, username, password,
port)
start_time = time()
retval, output = shell.command(command)
execution_time = time() - start_time
shell.disconnect()
msg = """
Command: %s
Return value: %s
Duration: %s seconds
Output:
%s""" % (command, retval, execution_time, output)
if retval:
notify_user(
user, "Deployment script failed for machine %s (%s)" %
(node.name, node.id), msg)
amqp_log(
"Deployment script failed for user %s machine %s (%s): %s"
% (user, node.name, node.id, msg))
else:
notify_user(
user, "Deployment script succeeded for machine %s (%s)" %
(node.name, node.id), msg)
amqp_log(
"Deployment script succeeded for user %s machine %s (%s): %s"
% (user, node.name, node.id, msg))
except ServiceUnavailableError as exc:
raise self.retry(exc=exc, countdown=60, max_retries=5)
except Exception as exc:
if str(exc).startswith('Retry'):
return
amqp_log(
"Deployment script failed for machine %s in backend %s by user %s after 5 retries: %s"
% (node.id, backend_id, email, repr(exc)))
notify_user(
user, "Deployment script failed for machine %s after 5 retries" %
node.id)
notify_admin(
"Deployment script failed for machine %s in backend %s by user %s after 5 retries"
% (node.id, backend_id, email), repr(exc))
def openstack_post_create_steps(self,
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
username,
password,
public_key,
script_id='',
script_params='',
job_id=None,
hostname='',
plugins=None,
post_script_id='',
post_script_params='',
networks=[]):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
conn = connect_provider(user.clouds[cloud_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
post_deploy_steps.delay(
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
script_id=script_id,
script_params=script_params,
job_id=job_id,
hostname=hostname,
plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
else:
try:
created_floating_ips = []
for network in networks['public']:
created_floating_ips += [
floating_ip for floating_ip in network['floating_ips']
]
# From the already created floating ips try to find one that is not associated to a node
unassociated_floating_ip = None
for ip in created_floating_ips:
if not ip['node_id']:
unassociated_floating_ip = ip
break
# Find the ports which are associated to the machine (e.g. the ports of the private ips)
# and use one to associate a floating ip
ports = conn.ex_list_ports()
machine_port_id = None
for port in ports:
if port.get('device_id') == node.id:
machine_port_id = port.get('id')
break
if unassociated_floating_ip:
log.info("Associating floating ip with machine: %s" %
node.id)
ip = conn.ex_associate_floating_ip_to_node(
unassociated_floating_ip['id'], machine_port_id)
else:
# Find the external network
log.info(
"Create and associating floating ip with machine: %s" %
node.id)
ext_net_id = networks['public'][0]['id']
ip = conn.ex_create_floating_ip(ext_net_id,
machine_port_id)
post_deploy_steps.delay(
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
script_id=script_id,
script_params=script_params,
job_id=job_id,
hostname=hostname,
plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except:
raise self.retry(exc=Exception(), max_retries=20)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def hpcloud_post_create_steps(self,
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
username,
password,
public_key,
script_id='',
script_params='',
job_id=None,
hostname='',
plugins=None,
post_script_id='',
post_script_params=''):
from mist.io.methods import connect_provider
user = user_from_email(email)
try:
conn = connect_provider(user.clouds[cloud_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and node.state == 0 and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
post_deploy_steps.delay(
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
script_id=script_id,
script_params=script_params,
job_id=job_id,
hostname=hostname,
plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
else:
try:
available_networks = conn.ex_list_networks()
external_networks = [
net for net in available_networks if net.router_external
]
if external_networks:
ext_net_id = external_networks[0].id
else:
ext_net_id = ""
ports = conn.ex_list_ports()
port = [
port for port in ports
if port.get("device_id", "") == node.id
][0]
ip = conn.ex_create_floating_ip(ext_net_id, port['id'])
post_deploy_steps.delay(
email,
cloud_id,
machine_id,
monitoring,
command,
key_id,
script_id=script_id,
script_params=script_params,
job_id=job_id,
hostname=hostname,
plugins=plugins,
post_script_id=post_script_id,
post_script_params=post_script_params,
)
except:
raise self.retry(exc=Exception(), max_retries=20)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
def post_deploy_steps(self,
email,
cloud_id,
machine_id,
monitoring,
command,
key_id=None,
username=None,
password=None,
port=22,
script_id='',
script_params='',
job_id=None,
hostname='',
plugins=None,
post_script_id='',
post_script_params=''):
from mist.io.methods import connect_provider, probe_ssh_only
from mist.io.methods import notify_user, notify_admin
from mist.io.methods import create_dns_a_record
if multi_user:
from mist.core.methods import enable_monitoring
from mist.core.tasks import run_script
from mist.core.helpers import log_event
else:
from mist.io.methods import enable_monitoring
log_event = lambda *args, **kwargs: None
job_id = job_id or uuid.uuid4().hex
user = user_from_email(email)
tmp_log = lambda msg, *args: log.error('Post deploy: %s' % msg, *args)
tmp_log('Entering post deploy steps for %s %s %s', user.email, cloud_id,
machine_id)
try:
# find the node we're looking for and get its hostname
node = None
try:
conn = connect_provider(user.clouds[cloud_id])
nodes = conn.list_nodes() # TODO: use cache
for n in nodes:
if n.id == machine_id:
node = n
break
tmp_log('run list_machines')
except:
raise self.retry(exc=Exception(), countdown=10, max_retries=10)
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
tmp_log('ip not found, retrying')
raise self.retry(exc=Exception(), countdown=60, max_retries=20)
try:
from mist.io.shell import Shell
shell = Shell(host)
# connect with ssh even if no command, to create association
# to be able to enable monitoring
tmp_log('attempting to connect to shell')
key_id, ssh_user = shell.autoconfigure(user, cloud_id, node.id,
key_id, username, password,
port)
tmp_log('connected to shell')
result = probe_ssh_only(user,
cloud_id,
machine_id,
host=None,
key_id=key_id,
ssh_user=ssh_user,
shell=shell)
log_dict = {
'email': email,
'event_type': 'job',
'cloud_id': cloud_id,
'machine_id': machine_id,
'job_id': job_id,
'host': host,
'key_id': key_id,
'ssh_user': ssh_user,
}
log_event(action='probe', result=result, **log_dict)
cloud = user.clouds[cloud_id]
msg = "Cloud:\n Name: %s\n Id: %s\n" % (cloud.title, cloud_id)
msg += "Machine:\n Name: %s\n Id: %s\n" % (node.name, node.id)
if hostname:
try:
record = create_dns_a_record(user, hostname, host)
hostname = '.'.join((record.name, record.zone.domain))
log_event(action='create_dns_a_record',
hostname=hostname,
**log_dict)
except Exception as exc:
log_event(action='create_dns_a_record',
error=str(exc),
**log_dict)
error = False
if script_id and multi_user:
tmp_log('will run script_id %s', script_id)
ret = run_script.run(user.email,
script_id,
cloud_id,
machine_id,
params=script_params,
host=host,
job_id=job_id)
error = ret['error']
tmp_log('executed script_id %s', script_id)
elif command:
tmp_log('will run command %s', command)
log_event(action='deployment_script_started',
command=command,
**log_dict)
start_time = time()
retval, output = shell.command(command)
tmp_log('executed command %s', command)
execution_time = time() - start_time
output = output.decode('utf-8', 'ignore')
title = "Deployment script %s" % ('failed'
if retval else 'succeeded')
error = retval > 0
notify_user(user,
title,
cloud_id=cloud_id,
machine_id=machine_id,
machine_name=node.name,
command=command,
output=output,
duration=execution_time,
retval=retval,
error=retval > 0)
log_event(action='deployment_script_finished',
error=retval > 0,
return_value=retval,
command=command,
stdout=output,
**log_dict)
shell.disconnect()
if monitoring:
try:
enable_monitoring(
user,
cloud_id,
node.id,
name=node.name,
dns_name=node.extra.get('dns_name', ''),
public_ips=ips,
no_ssh=False,
dry=False,
job_id=job_id,
plugins=plugins,
deploy_async=False,
)
except Exception as e:
print repr(e)
error = True
notify_user(
user,
"Enable monitoring failed for machine %s" % machine_id,
repr(e))
notify_admin(
'Enable monitoring on creation failed for user %s machine %s: %r'
% (email, machine_id, e))
log_event(action='enable_monitoring_failed',
error=repr(e),
**log_dict)
if post_script_id and multi_user:
tmp_log('will run post_script_id %s', post_script_id)
ret = run_script.run(
user.email,
post_script_id,
cloud_id,
machine_id,
params=post_script_params,
host=host,
job_id=job_id,
action_prefix='post_',
)
error = ret['error']
tmp_log('executed post_script_id %s', script_id)
log_event(action='post_deploy_finished', error=error, **log_dict)
except (ServiceUnavailableError, SSHException) as exc:
tmp_log(repr(exc))
raise self.retry(exc=exc, countdown=60, max_retries=15)
except Exception as exc:
tmp_log(repr(exc))
if str(exc).startswith('Retry'):
raise
notify_user(user,
"Deployment script failed for machine %s" % machine_id)
notify_admin(
"Deployment script failed for machine %s in cloud %s by user %s" %
(machine_id, cloud_id, email), repr(exc))
log_event(email=email,
event_type='job',
action='post_deploy_finished',
cloud_id=cloud_id,
machine_id=machine_id,
enable_monitoring=bool(monitoring),
command=command,
error="Couldn't connect to run post deploy steps.",
job_id=job_id)
def post_deploy_steps(self, email, backend_id, machine_id, monitoring, command,
key_id=None, username=None, password=None, port=22):
from mist.io.methods import ssh_command, connect_provider, enable_monitoring
from mist.io.methods import notify_user, notify_admin
if multi_user:
from mist.core.methods import enable_monitoring
from mist.core.helpers import log_event
else:
from mist.io.methods import enable_monitoring
log_event = lambda *args, **kwargs: None
user = user_from_email(email)
try:
# find the node we're looking for and get its hostname
conn = connect_provider(user.backends[backend_id])
nodes = conn.list_nodes()
node = None
for n in nodes:
if n.id == machine_id:
node = n
break
if node and len(node.public_ips):
# filter out IPv6 addresses
ips = filter(lambda ip: ':' not in ip, node.public_ips)
host = ips[0]
else:
raise self.retry(exc=Exception(), countdown=120, max_retries=5)
try:
from mist.io.shell import Shell
shell = Shell(host)
# connect with ssh even if no command, to create association
# to be able to enable monitoring
key_id, ssh_user = shell.autoconfigure(
user, backend_id, node.id, key_id, username, password, port
)
backend = user.backends[backend_id]
msg = "Backend:\n Name: %s\n Id: %s\n" % (backend.title,
backend_id)
msg += "Machine:\n Name: %s\n Id: %s\n" % (node.name,
node.id)
if command:
log_dict = {
'email': email,
'event_type': 'job',
'backend_id': backend_id,
'machine_id': machine_id,
'job_id': uuid.uuid4().hex,
'command': command,
'host': host,
'key_id': key_id,
'ssh_user': ssh_user,
}
log_event(action='deployment_script_started', **log_dict)
start_time = time()
retval, output = shell.command(command)
execution_time = time() - start_time
output = output.decode('utf-8','ignore')
title = "Deployment script %s" % ('failed' if retval
else 'succeeded')
notify_user(user, title,
backend_id=backend_id,
machine_id=machine_id,
machine_name=node.name,
command=command,
output=output,
duration=execution_time,
retval=retval,
error=retval > 0)
log_event(action='deployment_script_finished',
error=retval > 0,
return_value=retval,
stdout=output,
**log_dict)
shell.disconnect()
if monitoring:
try:
enable_monitoring(user, backend_id, node.id,
name=node.name, dns_name=node.extra.get('dns_name',''),
public_ips=ips, no_ssh=False, dry=False,
)
except Exception as e:
print repr(e)
notify_user(user, "Enable monitoring failed for machine %s (%s)" % (node.name, node.id), repr(e))
notify_admin('Enable monitoring on creation failed for user %s machine %s: %r' % (email, node.name, e))
except (ServiceUnavailableError, SSHException) as exc:
raise self.retry(exc=exc, countdown=60, max_retries=5)
except Exception as exc:
if str(exc).startswith('Retry'):
raise
notify_user(user, "Deployment script failed for machine %s after 5 retries" % node.id)
notify_admin("Deployment script failed for machine %s in backend %s by user %s after 5 retries" % (node.id, backend_id, email), repr(exc))
log_event(
email=email,
event_type='job',
action='deployment_script_failed',
backend_id=backend_id,
machine_id=machine_id,
enable_monitoring=bool(monitoring),
command=command,
error="Couldn't connect to run post deploy steps (5 attempts).",
)