diff -uNr a/fence/agents/compute/fence_compute.py b/fence/agents/compute/fence_compute.py
--- a/fence/agents/compute/fence_compute.py 2017-09-27 15:01:34.974642469 +0200
+++ b/fence/agents/compute/fence_compute.py 2017-09-27 15:24:57.482819900 +0200
@@ -18,173 +18,115 @@
#END_VERSION_GENERATION
override_status = ""
-nova = None
EVACUABLE_TAG = "evacuable"
TRUE_TAGS = ['true']
-def get_power_status(_, options):
- global override_status
-
- status = "unknown"
- logging.debug("get action: " + options["--action"])
+def get_power_status(connection, options):
if len(override_status):
logging.debug("Pretending we're " + override_status)
return override_status
- if nova:
+ status = "unknown"
+ logging.debug("get action: " + options["--action"])
+
+ if connection:
try:
- services = nova.services.list(host=options["--plug"])
+ services = connection.services.list(host=options["--plug"], binary="nova-compute")
for service in services:
- logging.debug("Status of %s is %s" % (service.binary, service.state))
- if service.binary == "nova-compute":
- if service.state == "up":
- status = "on"
- elif service.state == "down":
- status = "off"
- else:
- logging.debug("Unknown status detected from nova: " + service.state)
- break
+ logging.debug("Status of %s on %s is %s, %s" % (service.binary, options["--plug"], service.state, service.status))
+ if service.state == "up" and service.status == "enabled":
+ # Up and operational
+ status = "on"
+
+ elif service.state == "down" and service.status == "disabled":
+ # Down and fenced
+ status = "off"
+
+ elif service.state == "down":
+ # Down and requires fencing
+ status = "failed"
+
+ elif service.state == "up":
+ # Up and requires unfencing
+ status = "running"
+ else:
+ logging.warning("Unknown status detected from nova for %s: %s, %s" % (options["--plug"], service.state, service.status))
+ status = "%s %s" % (service.state, service.status)
+ break
except requests.exception.ConnectionError as err:
logging.warning("Nova connection failed: " + str(err))
+ logging.debug("Final status of %s is %s" % (options["--plug"], status))
return status
-# NOTE(sbauza); We mimic the host-evacuate module since it's only a contrib
-# module which is not stable
-def _server_evacuate(server, on_shared_storage):
- success = False
- error_message = ""
- try:
- logging.debug("Resurrecting instance: %s" % server)
- (response, dictionary) = nova.servers.evacuate(server=server, on_shared_storage=on_shared_storage)
-
- if response == None:
- error_message = "No response while evacuating instance"
- elif response.status_code == 200:
- success = True
- error_message = response.reason
- else:
- error_message = response.reason
-
- except Exception as e:
- error_message = "Error while evacuating instance: %s" % e
-
- return {
- "uuid": server,
- "accepted": success,
- "reason": error_message,
- }
-
-def _is_server_evacuable(server, evac_flavors, evac_images):
- if server.flavor.get('id') in evac_flavors:
- return True
- if server.image.get('id') in evac_images:
- return True
- logging.debug("Instance %s is not evacuable" % server.image.get('id'))
- return False
-
-def _get_evacuable_flavors():
- result = []
- flavors = nova.flavors.list()
- # Since the detailed view for all flavors doesn't provide the extra specs,
- # we need to call each of the flavor to get them.
- for flavor in flavors:
- tag = flavor.get_keys().get(EVACUABLE_TAG)
- if tag and tag.strip().lower() in TRUE_TAGS:
- result.append(flavor.id)
- return result
-
-def _get_evacuable_images():
- result = []
- images = nova.images.list(detailed=True)
- for image in images:
- if hasattr(image, 'metadata'):
- tag = image.metadata.get(EVACUABLE_TAG)
- if tag and tag.strip().lower() in TRUE_TAGS:
- result.append(image.id)
- return result
-
-def _host_evacuate(options):
- result = True
- images = _get_evacuable_images()
- flavors = _get_evacuable_flavors()
- servers = nova.servers.list(search_opts={'host': options["--plug"], 'all_tenants': 1 })
-
- if options["--instance-filtering"] == "False":
- logging.debug("Not evacuating anything")
- evacuables = []
- elif len(flavors) or len(images):
- logging.debug("Filtering images and flavors: %s %s" % (repr(flavors), repr(images)))
- # Identify all evacuable servers
- logging.debug("Checking %s" % repr(servers))
- evacuables = [server for server in servers
- if _is_server_evacuable(server, flavors, images)]
- logging.debug("Evacuating %s" % repr(evacuables))
- else:
- logging.debug("Evacuating all images and flavors")
- evacuables = servers
-
- if options["--no-shared-storage"] != "False":
- on_shared_storage = False
- else:
- on_shared_storage = True
-
- for server in evacuables:
- logging.debug("Processing %s" % server)
- if hasattr(server, 'id'):
- response = _server_evacuate(server.id, on_shared_storage)
- if response["accepted"]:
- logging.debug("Evacuated %s from %s: %s" %
- (response["uuid"], options["--plug"], response["reason"]))
- else:
- logging.error("Evacuation of %s on %s failed: %s" %
- (response["uuid"], options["--plug"], response["reason"]))
- result = False
- else:
- logging.error("Could not evacuate instance: %s" % server.to_dict())
- # Should a malformed instance result in a failed evacuation?
- # result = False
- return result
+def get_power_status_simple(connection, options):
+ status = get_power_status(connection, options)
+ if status in [ "off" ]:
+ return status
+ return "on"
def set_attrd_status(host, status, options):
logging.debug("Setting fencing status for %s to %s" % (host, status))
run_command(options, "attrd_updater -p -n evacuate -Q -N %s -U %s" % (host, status))
-def set_power_status(_, options):
- global override_status
-
- override_status = ""
- logging.debug("set action: " + options["--action"])
+def get_attrd_status(host, options):
+ (status, pipe_stdout, pipe_stderr) = run_command(options, "attrd_updater -p -n evacuate -Q -N %s" % (host))
+ fields = pipe_stdout.split('"')
+ if len(fields) > 6:
+ return fields[5]
+ logging.debug("Got %s: o:%s e:%s n:%d" % (status, pipe_stdout, pipe_stderr, len(fields)))
+ return ""
+
+def set_power_status_on(connection, options):
+ # Wait for any evacuations to complete
+ while True:
+ current = get_attrd_status(options["--plug"], options)
+ if current in ["no", ""]:
+ logging.info("Evacuation complete for: %s '%s'" % (options["--plug"], current))
+ break
+ else:
+ logging.info("Waiting for %s to complete evacuations: %s" % (options["--plug"], current))
+ time.sleep(2)
- if not nova:
- return
+ status = get_power_status(connection, options)
+ # Should we do it for 'failed' too?
+ if status in [ "off", "running", "failed" ]:
+ try:
+ # Forcing the host back up
+ logging.info("Forcing nova-compute back up on "+options["--plug"])
+ connection.services.force_down(options["--plug"], "nova-compute", force_down=False)
+ logging.info("Forced nova-compute back up on "+options["--plug"])
+ except Exception as e:
+ # In theory, if force_down=False fails, that's for the exact
+ # same possible reasons that below with force_down=True
+ # eg. either an incompatible version or an old client.
+ # Since it's about forcing back to a default value, there is
+ # no real worries to just consider it's still okay even if the
+ # command failed
+ logging.warn("Exception from attempt to force "
+ "host back up via nova API: "
+ "%s: %s" % (e.__class__.__name__, e))
+
+ # Forcing the service back up in case it was disabled
+ logging.info("Enabling nova-compute on "+options["--plug"])
+ connection.services.enable(options["--plug"], 'nova-compute')
- if options["--action"] == "on":
- if get_power_status(_, options) != "on":
- # Forcing the service back up in case it was disabled
- nova.services.enable(options["--plug"], 'nova-compute')
- try:
- # Forcing the host back up
- nova.services.force_down(
- options["--plug"], "nova-compute", force_down=False)
- except Exception as e:
- # In theory, if force_down=False fails, that's for the exact
- # same possible reasons that below with force_down=True
- # eg. either an incompatible version or an old client.
- # Since it's about forcing back to a default value, there is
- # no real worries to just consider it's still okay even if the
- # command failed
- logging.info("Exception from attempt to force "
- "host back up via nova API: "
- "%s: %s" % (e.__class__.__name__, e))
- else:
- # Pretend we're 'on' so that the fencing library doesn't loop forever waiting for the node to boot
- override_status = "on"
+ # Pretend we're 'on' so that the fencing library doesn't loop forever waiting for the node to boot
+ override_status = "on"
+ elif status not in ["on"]:
+ # Not safe to unfence, don't waste time looping to see if the status changes to "on"
+ options["--power-timeout"] = "0"
+
+def set_power_status_off(connection, options):
+ status = get_power_status(connection, options)
+ if status in [ "off" ]:
return
+ connection.services.disable(options["--plug"], 'nova-compute')
try:
- nova.services.force_down(
+ # Until 2.53
+ connection.services.force_down(
options["--plug"], "nova-compute", force_down=True)
except Exception as e:
# Something went wrong when we tried to force the host down.
@@ -198,7 +140,7 @@
"%s: %s" % (e.__class__.__name__, e))
# need to wait for nova to update its internal status or we
# cannot call host-evacuate
- while get_power_status(_, options) != "off":
+ while get_power_status(connection, options) not in ["off"]:
# Loop forever if need be.
#
# Some callers (such as Pacemaker) will have a timer
@@ -206,47 +148,55 @@
logging.debug("Waiting for nova to update its internal state for %s" % options["--plug"])
time.sleep(1)
- if not _host_evacuate(options):
- sys.exit(1)
+ set_attrd_status(options["--plug"], "yes", options)
+
+def set_power_status(connection, options):
+ global override_status
- return
+ override_status = ""
+ logging.debug("set action: " + options["--action"])
+
+ if not connection:
+ return
+ if options["--action"] in ["off", "reboot"]:
+ set_power_status_off(connection, options)
+ else:
+ set_power_status_on(connection, options)
+ logging.debug("set action passed: " + options["--action"])
+ sys.exit(0)
-def fix_domain(options):
+def fix_domain(connection, options):
domains = {}
last_domain = None
- if nova:
+ if connection:
# Find it in nova
- hypervisors = nova.hypervisors.list()
- for hypervisor in hypervisors:
- shorthost = hypervisor.hypervisor_hostname.split('.')[0]
+ services = connection.services.list(binary="nova-compute")
+ for service in services:
+ shorthost = service.host.split('.')[0]
- if shorthost == hypervisor.hypervisor_hostname:
+ if shorthost == service.host:
# Nova is not using FQDN
calculated = ""
else:
# Compute nodes are named as FQDN, strip off the hostname
- calculated = hypervisor.hypervisor_hostname.replace(shorthost+".", "")
-
- domains[calculated] = shorthost
+ calculated = service.host.replace(shorthost+".", "")
if calculated == last_domain:
# Avoid complaining for each compute node with the same name
# One hopes they don't appear interleaved as A.com B.com A.com B.com
- logging.debug("Calculated the same domain from: %s" % hypervisor.hypervisor_hostname)
+ logging.debug("Calculated the same domain from: %s" % service.host)
+ continue
- elif "--domain" in options and options["--domain"] == calculated:
- # Supplied domain name is valid
- return
+ domains[calculated] = service.host
+ last_domain = calculated
- elif "--domain" in options:
+ if "--domain" in options and options["--domain"] != calculated:
# Warn in case nova isn't available at some point
logging.warning("Supplied domain '%s' does not match the one calculated from: %s"
- % (options["--domain"], hypervisor.hypervisor_hostname))
-
- last_domain = calculated
+ % (options["--domain"], service.host))
if len(domains) == 0 and "--domain" not in options:
logging.error("Could not calculate the domain names used by compute nodes in nova")
@@ -254,9 +204,9 @@
elif len(domains) == 1 and "--domain" not in options:
options["--domain"] = last_domain
- elif len(domains) == 1:
- logging.error("Overriding supplied domain '%s' does not match the one calculated from: %s"
- % (options["--domain"], hypervisor.hypervisor_hostname))
+ elif len(domains) == 1 and options["--domain"] != last_domain:
+ logging.error("Overriding supplied domain '%s' as it does not match the one calculated from: %s"
+ % (options["--domain"], domains[last_domain]))
options["--domain"] = last_domain
elif len(domains) > 1:
@@ -264,47 +214,49 @@
% (options["--domain"], repr(domains)))
sys.exit(1)
-def fix_plug_name(options):
+ return last_domain
+
+def fix_plug_name(connection, options):
if options["--action"] == "list":
return
if "--plug" not in options:
return
- fix_domain(options)
- short_plug = options["--plug"].split('.')[0]
- logging.debug("Checking target '%s' against calculated domain '%s'"% (options["--plug"], options["--domain"]))
-
- if "--domain" not in options:
+ calculated = fix_domain(connection, options)
+ if calculated is None or "--domain" not in options:
# Nothing supplied and nova not available... what to do... nothing
return
- elif options["--domain"] == "":
+ short_plug = options["--plug"].split('.')[0]
+ logging.debug("Checking target '%s' against calculated domain '%s'"% (options["--plug"], calculated))
+
+ if options["--domain"] == "":
# Ensure any domain is stripped off since nova isn't using FQDN
options["--plug"] = short_plug
- elif options["--domain"] in options["--plug"]:
- # Plug already contains the domain, don't re-add
+ elif options["--plug"].endswith(options["--domain"]):
+ # Plug already uses the domain, don't re-add
return
else:
# Add the domain to the plug
options["--plug"] = short_plug + "." + options["--domain"]
-def get_plugs_list(_, options):
+def get_plugs_list(connection, options):
result = {}
- if nova:
- hypervisors = nova.hypervisors.list()
- for hypervisor in hypervisors:
- longhost = hypervisor.hypervisor_hostname
+ if connection:
+ services = connection.services.list(binary="nova-compute")
+ for service in services:
+ longhost = service.host
shorthost = longhost.split('.')[0]
result[longhost] = ("", None)
result[shorthost] = ("", None)
return result
def create_nova_connection(options):
- global nova
+ nova = None
try:
from novaclient import client
@@ -330,41 +282,42 @@
if clientargs:
# OSP < 11
# ArgSpec(args=['version', 'username', 'password', 'project_id', 'auth_url'],
- # varargs=None,
- # keywords='kwargs', defaults=(None, None, None, None))
+ # varargs=None,
+ # keywords='kwargs', defaults=(None, None, None, None))
nova = client.Client(version,
- options["--username"],
- options["--password"],
- options["--tenant-name"],
- options["--auth-url"],
- insecure=options["--insecure"],
- region_name=options["--region-name"],
- endpoint_type=options["--endpoint-type"],
- http_log_debug=options.has_key("--verbose"))
+ options["--username"],
+ options["--password"],
+ options["--tenant-name"],
+ options["--auth-url"],
+ insecure=options["--insecure"],
+ region_name=options["--region-name"],
+ endpoint_type=options["--endpoint-type"],
+ http_log_debug=options.has_key("--verbose"))
else:
# OSP >= 11
# ArgSpec(args=['version'], varargs='args', keywords='kwargs', defaults=None)
nova = client.Client(version,
- username=options["--username"],
- password=options["--password"],
- tenant_name=options["--tenant-name"],
- auth_url=options["--auth-url"],
- insecure=options["--insecure"],
- region_name=options["--region-name"],
- endpoint_type=options["--endpoint-type"],
- http_log_debug=options.has_key("--verbose"))
+ username=options["--username"],
+ password=options["--password"],
+ tenant_name=options["--tenant-name"],
+ auth_url=options["--auth-url"],
+ insecure=options["--insecure"],
+ region_name=options["--region-name"],
+ endpoint_type=options["--endpoint-type"],
+ http_log_debug=options.has_key("--verbose"))
try:
nova.hypervisors.list()
- return
+ return nova
except NotAcceptable as e:
logging.warning(e)
except Exception as e:
logging.warning("Nova connection failed. %s: %s" % (e.__class__.__name__, e))
-
+
logging.warning("Couldn't obtain a supported connection to nova, tried: %s\n" % repr(versions))
+ return None
def define_new_opts():
all_opt["endpoint_type"] = {
@@ -448,11 +401,23 @@
"order": 5,
}
+def set_multi_power_fn(connection, options, set_power_fn, get_power_fn, retry_attempts=1):
+ for _ in range(retry_attempts):
+ set_power_fn(connection, options)
+ time.sleep(int(options["--power-wait"]))
+
+ for _ in range(int(options["--power-timeout"])):
+ if get_power_fn(connection, options) != options["--action"]:
+ time.sleep(1)
+ else:
+ return True
+ return False
+
def main():
global override_status
atexit.register(atexit_handler)
- device_opt = ["login", "passwd", "tenant_name", "auth_url", "fabric_fencing",
+ device_opt = ["login", "passwd", "tenant_name", "auth_url", "fabric_fencing",
"no_login", "no_password", "port", "domain", "no_shared_storage", "endpoint_type",
"record_only", "instance_filtering", "insecure", "region_name"]
define_new_opts()
@@ -472,30 +437,28 @@
run_delay(options)
- create_nova_connection(options)
+ logging.debug("Running "+options["--action"])
+ connection = create_nova_connection(options)
- fix_plug_name(options)
- if options["--record-only"] in [ "1", "True", "true", "Yes", "yes"]:
- if options["--action"] == "on":
- set_attrd_status(options["--plug"], "no", options)
- sys.exit(0)
-
- elif options["--action"] in ["off", "reboot"]:
- set_attrd_status(options["--plug"], "yes", options)
- sys.exit(0)
+ if options["--action"] in ["off", "on", "reboot", "status"]:
+ fix_plug_name(connection, options)
- elif options["--action"] in ["monitor", "status"]:
- sys.exit(0)
- if options["--action"] in ["off", "reboot"]:
- # Pretend we're 'on' so that the fencing library will always call set_power_status(off)
- override_status = "on"
-
- if options["--action"] == "on":
- # Pretend we're 'off' so that the fencing library will always call set_power_status(on)
- override_status = "off"
+ if options["--action"] in ["reboot"]:
+ options["--action"]="off"
+
+ if options["--action"] in ["off", "on"]:
+ # No status first, call our own version
+ result = not set_multi_power_fn(connection, options, set_power_status, get_power_status_simple,
+ 1 + int(options["--retry-on"]))
+ elif options["--action"] in ["monitor"]:
+ result = 0
+ else:
+ result = fence_action(connection, options, set_power_status, get_power_status_simple, get_plugs_list, None)
- result = fence_action(None, options, set_power_status, get_power_status, get_plugs_list, None)
+ logging.debug("Result for "+options["--action"]+": "+repr(result))
+ if result == None:
+ result = 0
sys.exit(result)
if __name__ == "__main__":
diff -uNr a/fence/agents/compute/fence_evacuate.py b/fence/agents/compute/fence_evacuate.py
--- a/fence/agents/compute/fence_evacuate.py 1970-01-01 01:00:00.000000000 +0100
+++ b/fence/agents/compute/fence_evacuate.py 2017-09-27 15:25:54.234304769 +0200
@@ -0,0 +1,366 @@
+#!/usr/bin/python -tt
+
+import sys
+import time
+import atexit
+import logging
+import inspect
+import requests.exceptions
+
+sys.path.append("@FENCEAGENTSLIBDIR@")
+from fencing import *
+from fencing import fail_usage, is_executable, run_command, run_delay
+
+EVACUABLE_TAG = "evacuable"
+TRUE_TAGS = ['true']
+
+def get_power_status(connection, options):
+
+ status = "unknown"
+ logging.debug("get action: " + options["--action"])
+
+ if connection:
+ try:
+ services = connection.services.list(host=options["--plug"], binary="nova-compute")
+ for service in services:
+ logging.debug("Status of %s is %s, %s" % (service.binary, service.state, service.status))
+ if service.state == "up" and service.status == "enabled":
+ # Up and operational
+ status = "on"
+
+ elif service.state == "down" and service.status == "disabled":
+ # Down and fenced
+ status = "off"
+
+ elif service.state == "down":
+ # Down and requires fencing
+ status = "failed"
+
+ elif service.state == "up":
+ # Up and requires unfencing
+ status = "running"
+ else:
+ logging.warning("Unknown status detected from nova for %s: %s, %s" % (options["--plug"], service.state, service.status))
+ status = "%s %s" % (service.state, service.status)
+ break
+ except requests.exception.ConnectionError as err:
+ logging.warning("Nova connection failed: " + str(err))
+ return status
+
+# NOTE(sbauza); We mimic the host-evacuate module since it's only a contrib
+# module which is not stable
+def _server_evacuate(connection, server, on_shared_storage):
+ success = False
+ error_message = ""
+ try:
+ logging.debug("Resurrecting instance: %s" % server)
+ (response, dictionary) = connection.servers.evacuate(server=server, on_shared_storage=on_shared_storage)
+
+ if response == None:
+ error_message = "No response while evacuating instance"
+ elif response.status_code == 200:
+ success = True
+ error_message = response.reason
+ else:
+ error_message = response.reason
+
+ except Exception as e:
+ error_message = "Error while evacuating instance: %s" % e
+
+ return {
+ "uuid": server,
+ "accepted": success,
+ "reason": error_message,
+ }
+
+def _is_server_evacuable(server, evac_flavors, evac_images):
+ if server.flavor.get('id') in evac_flavors:
+ return True
+ if hasattr(server.image, 'get'):
+ if server.image.get('id') in evac_images:
+ return True
+ logging.debug("Instance %s is not evacuable" % server.image.get('id'))
+ return False
+
+def _get_evacuable_flavors(connection):
+ result = []
+ flavors = connection.flavors.list()
+ # Since the detailed view for all flavors doesn't provide the extra specs,
+ # we need to call each of the flavor to get them.
+ for flavor in flavors:
+ tag = flavor.get_keys().get(EVACUABLE_TAG)
+ if tag and tag.strip().lower() in TRUE_TAGS:
+ result.append(flavor.id)
+ return result
+
+def _get_evacuable_images(connection):
+ result = []
+ images = []
+ if hasattr(connection, "images"):
+ images = connection.images.list(detailed=True)
+ elif hasattr(connection, "glance"):
+ # OSP12+
+ images = connection.glance.list()
+
+ for image in images:
+ if hasattr(image, 'metadata'):
+ tag = image.metadata.get(EVACUABLE_TAG)
+ if tag and tag.strip().lower() in TRUE_TAGS:
+ result.append(image.id)
+ elif hasattr(image, 'tags'):
+ # OSP12+
+ if EVACUABLE_TAG in image.tags:
+ result.append(image.id)
+ return result
+
+def _host_evacuate(connection, options):
+ result = True
+ images = _get_evacuable_images(connection)
+ flavors = _get_evacuable_flavors(connection)
+ servers = connection.servers.list(search_opts={'host': options["--plug"], 'all_tenants': 1 })
+
+ if options["--instance-filtering"] == "False":
+ logging.debug("Not evacuating anything")
+ evacuables = []
+ elif len(flavors) or len(images):
+ logging.debug("Filtering images and flavors: %s %s" % (repr(flavors), repr(images)))
+ # Identify all evacuable servers
+ logging.debug("Checking %s" % repr(servers))
+ evacuables = [server for server in servers
+ if _is_server_evacuable(server, flavors, images)]
+ logging.debug("Evacuating %s" % repr(evacuables))
+ else:
+ logging.debug("Evacuating all images and flavors")
+ evacuables = servers
+
+ if options["--no-shared-storage"] != "False":
+ on_shared_storage = False
+ else:
+ on_shared_storage = True
+
+ for server in evacuables:
+ logging.debug("Processing %s" % server)
+ if hasattr(server, 'id'):
+ response = _server_evacuate(connection, server.id, on_shared_storage)
+ if response["accepted"]:
+ logging.debug("Evacuated %s from %s: %s" %
+ (response["uuid"], options["--plug"], response["reason"]))
+ else:
+ logging.error("Evacuation of %s on %s failed: %s" %
+ (response["uuid"], options["--plug"], response["reason"]))
+ result = False
+ else:
+ logging.error("Could not evacuate instance: %s" % server.to_dict())
+ # Should a malformed instance result in a failed evacuation?
+ # result = False
+ return result
+
+def set_attrd_status(host, status, options):
+ logging.debug("Setting fencing status for %s to %s" % (host, status))
+ run_command(options, "attrd_updater -p -n evacuate -Q -N %s -U %s" % (host, status))
+
+def set_power_status(connection, options):
+ logging.debug("set action: " + options["--action"])
+
+ if not connection:
+ return
+
+ if options["--action"] == "off" and not _host_evacuate(options):
+ sys.exit(1)
+
+ sys.exit(0)
+
+def get_plugs_list(connection, options):
+ result = {}
+
+ if connection:
+ services = connection.services.list(binary="nova-compute")
+ for service in services:
+ longhost = service.host
+ shorthost = longhost.split('.')[0]
+ result[longhost] = ("", None)
+ result[shorthost] = ("", None)
+ return result
+
+def create_nova_connection(options):
+ nova = None
+
+ try:
+ from novaclient import client
+ from novaclient.exceptions import NotAcceptable
+ except ImportError:
+ fail_usage("Nova not found or not accessible")
+
+ versions = [ "2.11", "2" ]
+ for version in versions:
+ clientargs = inspect.getargspec(client.Client).varargs
+
+ # Some versions of Openstack prior to Ocata only
+ # supported positional arguments for username,
+ # password and tenant.
+ #
+ # Versions since Ocata only support named arguments.
+ #
+ # So we need to use introspection to figure out how to
+ # create a Nova client.
+ #
+ # Happy days
+ #
+ if clientargs:
+ # OSP < 11
+ # ArgSpec(args=['version', 'username', 'password', 'project_id', 'auth_url'],
+ # varargs=None,
+ # keywords='kwargs', defaults=(None, None, None, None))
+ nova = client.Client(version,
+ options["--username"],
+ options["--password"],
+ options["--tenant-name"],
+ options["--auth-url"],
+ insecure=options["--insecure"],
+ region_name=options["--region-name"],
+ endpoint_type=options["--endpoint-type"],
+ http_log_debug=options.has_key("--verbose"))
+ else:
+ # OSP >= 11
+ # ArgSpec(args=['version'], varargs='args', keywords='kwargs', defaults=None)
+ nova = client.Client(version,
+ username=options["--username"],
+ password=options["--password"],
+ tenant_name=options["--tenant-name"],
+ auth_url=options["--auth-url"],
+ insecure=options["--insecure"],
+ region_name=options["--region-name"],
+ endpoint_type=options["--endpoint-type"],
+ http_log_debug=options.has_key("--verbose"))
+
+ try:
+ nova.hypervisors.list()
+ return nova
+
+ except NotAcceptable as e:
+ logging.warning(e)
+
+ except Exception as e:
+ logging.warning("Nova connection failed. %s: %s" % (e.__class__.__name__, e))
+
+ logging.warning("Couldn't obtain a supported connection to nova, tried: %s\n" % repr(versions))
+ return None
+
+def define_new_opts():
+ all_opt["endpoint_type"] = {
+ "getopt" : "e:",
+ "longopt" : "endpoint-type",
+ "help" : "-e, --endpoint-type=[endpoint] Nova Endpoint type (publicURL, internalURL, adminURL)",
+ "required" : "0",
+ "shortdesc" : "Nova Endpoint type",
+ "default" : "internalURL",
+ "order": 1,
+ }
+ all_opt["tenant_name"] = {
+ "getopt" : "t:",
+ "longopt" : "tenant-name",
+ "help" : "-t, --tenant-name=[tenant] Keystone Admin Tenant",
+ "required" : "0",
+ "shortdesc" : "Keystone Admin Tenant",
+ "default" : "",
+ "order": 1,
+ }
+ all_opt["auth_url"] = {
+ "getopt" : "k:",
+ "longopt" : "auth-url",
+ "help" : "-k, --auth-url=[url] Keystone Admin Auth URL",
+ "required" : "0",
+ "shortdesc" : "Keystone Admin Auth URL",
+ "default" : "",
+ "order": 1,
+ }
+ all_opt["region_name"] = {
+ "getopt" : "",
+ "longopt" : "region-name",
+ "help" : "--region-name=[region] Region Name",
+ "required" : "0",
+ "shortdesc" : "Region Name",
+ "default" : "",
+ "order": 1,
+ }
+ all_opt["insecure"] = {
+ "getopt" : "",
+ "longopt" : "insecure",
+ "help" : "--insecure Explicitly allow agent to perform \"insecure\" TLS (https) requests",
+ "required" : "0",
+ "shortdesc" : "Allow Insecure TLS Requests",
+ "default" : "False",
+ "order": 2,
+ }
+ all_opt["domain"] = {
+ "getopt" : "d:",
+ "longopt" : "domain",
+ "help" : "-d, --domain=[string] DNS domain in which hosts live, useful when the cluster uses short names and nova uses FQDN",
+ "required" : "0",
+ "shortdesc" : "DNS domain in which hosts live",
+ "order": 5,
+ }
+ all_opt["instance_filtering"] = {
+ "getopt" : "",
+ "longopt" : "instance-filtering",
+ "help" : "--instance-filtering Allow instances created from images and flavors with evacuable=true to be evacuated (or all if no images/flavors have been tagged)",
+ "required" : "0",
+ "shortdesc" : "Allow instances to be evacuated",
+ "default" : "True",
+ "order": 5,
+ }
+ all_opt["no_shared_storage"] = {
+ "getopt" : "",
+ "longopt" : "no-shared-storage",
+ "help" : "--no-shared-storage Disable functionality for shared storage",
+ "required" : "0",
+ "shortdesc" : "Disable functionality for dealing with shared storage",
+ "default" : "False",
+ "order": 5,
+ }
+
+def main():
+ atexit.register(atexit_handler)
+
+ device_opt = ["login", "passwd", "tenant_name", "auth_url",
+ "no_login", "no_password", "port", "domain", "no_shared_storage", "endpoint_type",
+ "instance_filtering", "insecure", "region_name"]
+ define_new_opts()
+ all_opt["shell_timeout"]["default"] = "180"
+
+ options = check_input(device_opt, process_input(device_opt))
+
+ docs = {}
+ docs["shortdesc"] = "Fence agent for the automatic resurrection of OpenStack compute instances"
+ docs["longdesc"] = "Used to reschedule flagged instances"
+ docs["vendorurl"] = ""
+
+ show_docs(options, docs)
+
+ run_delay(options)
+
+ connection = create_nova_connection(options)
+
+ # Un-evacuating a server doesn't make sense
+ if options["--action"] in ["on"]:
+ logging.error("Action %s is not supported by this agent" % (options["--action"]))
+ sys.exit(1)
+
+ if options["--action"] in ["off", "reboot"]:
+ status = get_power_status(connection, options)
+ if status != "off":
+ logging.error("Cannot resurrect instances from %s in state '%s'" % (options["--plug"], status))
+ sys.exit(1)
+
+ elif not _host_evacuate(connection, options):
+ logging.error("Resurrection of instances from %s failed" % (options["--plug"]))
+ sys.exit(1)
+
+ logging.info("Resurrection of instances from %s complete" % (options["--plug"]))
+ sys.exit(0)
+
+ result = fence_action(connection, options, set_power_status, get_power_status, get_plugs_list, None)
+ sys.exit(result)
+
+if __name__ == "__main__":
+ main()
diff -uNr a/fence/agents/compute/Makefile.am b/fence/agents/compute/Makefile.am
--- a/fence/agents/compute/Makefile.am 2017-09-27 15:01:34.844643650 +0200
+++ b/fence/agents/compute/Makefile.am 2017-09-27 15:57:50.963839738 +0200
@@ -1,14 +1,14 @@
MAINTAINERCLEANFILES = Makefile.in
-TARGET = fence_compute
+TARGET = fence_compute fence_evacuate
-SRC = $(TARGET).py
+SRC = $(TARGET:=.py)
EXTRA_DIST = $(SRC)
sbin_SCRIPTS = $(TARGET)
-man_MANS = $(TARGET).8
+man_MANS = $(TARGET:=.8)
FENCE_TEST_ARGS = -l test -p test -n 1
diff -uNr a/tests/data/metadata/fence_evacuate.xml b/tests/data/metadata/fence_evacuate.xml
--- a/tests/data/metadata/fence_evacuate.xml 1970-01-01 01:00:00.000000000 +0100
+++ b/tests/data/metadata/fence_evacuate.xml 2017-09-27 15:28:10.978063549 +0200
@@ -0,0 +1,163 @@
+<?xml version="1.0" ?>
+<resource-agent name="fence_evacuate" shortdesc="Fence agent for the automatic resurrection of OpenStack compute instances" >
+<longdesc>Used to reschedule flagged instances</longdesc>
+<vendor-url></vendor-url>
+<parameters>
+ <parameter name="tenant_name" unique="0" required="0">
+ <getopt mixed="-t, --tenant-name=[tenant]" />
+ <content type="string" />
+ <shortdesc lang="en">Keystone Admin Tenant</shortdesc>
+ </parameter>
+ <parameter name="auth_url" unique="0" required="0">
+ <getopt mixed="-k, --auth-url=[url]" />
+ <content type="string" />
+ <shortdesc lang="en">Keystone Admin Auth URL</shortdesc>
+ </parameter>
+ <parameter name="port" unique="0" required="1" deprecated="1">
+ <getopt mixed="-n, --plug=[id]" />
+ <content type="string" />
+ <shortdesc lang="en">Physical plug number, name of virtual machine or UUID</shortdesc>
+ </parameter>
+ <parameter name="passwd_script" unique="0" required="0" deprecated="1">
+ <getopt mixed="-S, --password-script=[script]" />
+ <content type="string" />
+ <shortdesc lang="en">Script to retrieve password</shortdesc>
+ </parameter>
+ <parameter name="region_name" unique="0" required="0">
+ <getopt mixed="--region-name=[region]" />
+ <content type="boolean" />
+ <shortdesc lang="en">Region Name</shortdesc>
+ </parameter>
+ <parameter name="passwd" unique="0" required="0" deprecated="1">
+ <getopt mixed="-p, --password=[password]" />
+ <content type="string" />
+ <shortdesc lang="en">Login password or passphrase</shortdesc>
+ </parameter>
+ <parameter name="endpoint_type" unique="0" required="0">
+ <getopt mixed="-e, --endpoint-type=[endpoint]" />
+ <content type="string" default="internalURL" />
+ <shortdesc lang="en">Nova Endpoint type</shortdesc>
+ </parameter>
+ <parameter name="action" unique="0" required="1">
+ <getopt mixed="-o, --action=[action]" />
+ <content type="string" default="reboot" />
+ <shortdesc lang="en">Fencing Action</shortdesc>
+ </parameter>
+ <parameter name="login" unique="0" required="0" deprecated="1">
+ <getopt mixed="-l, --username=[name]" />
+ <content type="string" />
+ <shortdesc lang="en">Login Name</shortdesc>
+ </parameter>
+ <parameter name="plug" unique="0" required="1" obsoletes="port">
+ <getopt mixed="-n, --plug=[id]" />
+ <content type="string" />
+ <shortdesc lang="en">Physical plug number, name of virtual machine or UUID</shortdesc>
+ </parameter>
+ <parameter name="username" unique="0" required="0" obsoletes="login">
+ <getopt mixed="-l, --username=[name]" />
+ <content type="string" />
+ <shortdesc lang="en">Login Name</shortdesc>
+ </parameter>
+ <parameter name="password" unique="0" required="0" obsoletes="passwd">
+ <getopt mixed="-p, --password=[password]" />
+ <content type="string" />
+ <shortdesc lang="en">Login password or passphrase</shortdesc>
+ </parameter>
+ <parameter name="password_script" unique="0" required="0" obsoletes="passwd_script">
+ <getopt mixed="-S, --password-script=[script]" />
+ <content type="string" />
+ <shortdesc lang="en">Script to retrieve password</shortdesc>
+ </parameter>
+ <parameter name="insecure" unique="0" required="0">
+ <getopt mixed="--insecure" />
+ <content type="boolean" default="False" />
+ <shortdesc lang="en">Allow Insecure TLS Requests</shortdesc>
+ </parameter>
+ <parameter name="domain" unique="0" required="0">
+ <getopt mixed="-d, --domain=[string]" />
+ <content type="string" />
+ <shortdesc lang="en">DNS domain in which hosts live</shortdesc>
+ </parameter>
+ <parameter name="instance_filtering" unique="0" required="0">
+ <getopt mixed="--instance-filtering" />
+ <content type="boolean" default="True" />
+ <shortdesc lang="en">Allow instances to be evacuated</shortdesc>
+ </parameter>
+ <parameter name="no_shared_storage" unique="0" required="0">
+ <getopt mixed="--no-shared-storage" />
+ <content type="boolean" default="False" />
+ <shortdesc lang="en">Disable functionality for dealing with shared storage</shortdesc>
+ </parameter>
+ <parameter name="verbose" unique="0" required="0">
+ <getopt mixed="-v, --verbose" />
+ <content type="boolean" />
+ <shortdesc lang="en">Verbose mode</shortdesc>
+ </parameter>
+ <parameter name="debug" unique="0" required="0" deprecated="1">
+ <getopt mixed="-D, --debug-file=[debugfile]" />
+ <content type="string" />
+ <shortdesc lang="en">Write debug information to given file</shortdesc>
+ </parameter>
+ <parameter name="debug_file" unique="0" required="0" obsoletes="debug">
+ <getopt mixed="-D, --debug-file=[debugfile]" />
+ <content type="string" />
+ <shortdesc lang="en">Write debug information to given file</shortdesc>
+ </parameter>
+ <parameter name="version" unique="0" required="0">
+ <getopt mixed="-V, --version" />
+ <content type="boolean" />
+ <shortdesc lang="en">Display version information and exit</shortdesc>
+ </parameter>
+ <parameter name="help" unique="0" required="0">
+ <getopt mixed="-h, --help" />
+ <content type="boolean" />
+ <shortdesc lang="en">Display help and exit</shortdesc>
+ </parameter>
+ <parameter name="separator" unique="0" required="0">
+ <getopt mixed="-C, --separator=[char]" />
+ <content type="string" default="," />
+ <shortdesc lang="en">Separator for CSV created by operation list</shortdesc>
+ </parameter>
+ <parameter name="power_wait" unique="0" required="0">
+ <getopt mixed="--power-wait=[seconds]" />
+ <content type="second" default="0" />
+ <shortdesc lang="en">Wait X seconds after issuing ON/OFF</shortdesc>
+ </parameter>
+ <parameter name="login_timeout" unique="0" required="0">
+ <getopt mixed="--login-timeout=[seconds]" />
+ <content type="second" default="5" />
+ <shortdesc lang="en">Wait X seconds for cmd prompt after login</shortdesc>
+ </parameter>
+ <parameter name="delay" unique="0" required="0">
+ <getopt mixed="--delay=[seconds]" />
+ <content type="second" default="0" />
+ <shortdesc lang="en">Wait X seconds before fencing is started</shortdesc>
+ </parameter>
+ <parameter name="power_timeout" unique="0" required="0">
+ <getopt mixed="--power-timeout=[seconds]" />
+ <content type="second" default="20" />
+ <shortdesc lang="en">Test X seconds for status change after ON/OFF</shortdesc>
+ </parameter>
+ <parameter name="shell_timeout" unique="0" required="0">
+ <getopt mixed="--shell-timeout=[seconds]" />
+ <content type="second" default="180" />
+ <shortdesc lang="en">Wait X seconds for cmd prompt after issuing command</shortdesc>
+ </parameter>
+ <parameter name="retry_on" unique="0" required="0">
+ <getopt mixed="--retry-on=[attempts]" />
+ <content type="integer" default="1" />
+ <shortdesc lang="en">Count of attempts to retry power on</shortdesc>
+ </parameter>
+</parameters>
+<actions>
+ <action name="on" automatic="0"/>
+ <action name="off" />
+ <action name="reboot" />
+ <action name="status" />
+ <action name="list" />
+ <action name="list-status" />
+ <action name="monitor" />
+ <action name="metadata" />
+ <action name="validate-all" />
+</actions>
+</resource-agent>