Updated ping command to work with MCC/MOS
- updated Pinger class with inherited structure for Salt and Kube
- implemented DeamonSet handling in KubeApi interface
- implemented put-textfile and series of ConfigMap methods in KubeApi
- updated Pinger to use multiple --cidr commands at once
- update Summary section to be more informative and human readable
Change-Id: Iac18a619d0bb9a36a286a07f38aeba8f99a454ca
Related-PROD: PROD-36603
diff --git a/cfg_checker/modules/network/__init__.py b/cfg_checker/modules/network/__init__.py
index 714c170..4c95ef3 100644
--- a/cfg_checker/modules/network/__init__.py
+++ b/cfg_checker/modules/network/__init__.py
@@ -16,11 +16,15 @@
_class = checker.SaltNetworkChecker
elif strClassHint == "mapper":
_class = mapper.SaltNetworkMapper
+ elif strClassHint == "pinger":
+ _class = pinger.SaltNetworkPinger
elif _env == ENV_TYPE_KUBE:
if strClassHint == "checker":
_class = checker.KubeNetworkChecker
elif strClassHint == "mapper":
_class = mapper.KubeNetworkMapper
+ elif strClassHint == "pinger":
+ _class = pinger.KubeNetworkPinger
if not _class:
raise CheckerException(
"Unknown hint for selecting Network handler Class: '{}'".format(
@@ -73,12 +77,13 @@
net_ping_parser.add_argument(
'--cidr',
metavar='network_ping_cidr',
- help="Subnet CIDR to ping nodes in"
+ action="append",
+ help="Subnet CIDR to ping nodes in. Can be used multiple times"
)
net_ping_parser.add_argument(
'--mtu',
- metavar='network_ping_mtu',
- help="MTU size to use. Ping will be done for MTU - 20 - 8"
+ metavar='network_ping_mtu', default=64,
+ help="MTU size to use. Ping will be done for MTU - 20 - 8. Default: 64"
)
net_ping_parser.add_argument(
'--detailed',
@@ -240,12 +245,55 @@
# Checks if selected nodes are pingable
# with a desireble parameters: MTU, Frame, etc
# Check if there is supported env found
- _env = args_utils.check_supported_env(ENV_TYPE_SALT, args, config)
+ _env = args_utils.check_supported_env(
+ [ENV_TYPE_SALT, ENV_TYPE_KUBE],
+ args,
+ config
+ )
# Start command
+ # prepare parameters
if not args.cidr:
logger_cli.error("\n# Use mcp-check network list to get list of CIDRs")
_cidr = args_utils.get_arg(args, "cidr")
_skip, _skip_file = args_utils.get_skip_args(args)
+ _mtu = args_utils.get_arg(args, "mtu")
+ if isinstance(_mtu, str):
+ if _mtu == "auto":
+ logger_cli.info(
+ "# Supplied MTU is 'auto'. Network's MTU value will be used"
+ )
+ _mtu = 0
+ else:
+ raise CheckerException(
+ "Bad MTU value supplied: '{}'.\n"
+ "Allowed values are: 'auto'/0 (autodetect), 64 - 9100".format(
+ _mtu
+ )
+ )
+ elif isinstance(_mtu, int):
+ if _mtu < 64 and _mtu > 0:
+ logger_cli.warn(
+ "WARNING: Supplied MTU value is low: '{}'"
+ "Defaulting to '64'".format(_mtu)
+ )
+ _mtu = 64
+ elif _mtu == 0:
+ logger_cli.info(
+ "# MTU set to '0'. Network's MTU value will be used."
+ )
+ elif _mtu > 9100:
+ logger_cli.warn(
+ "WARNING: Supplied MTU value is >9100: '{}'"
+ "Defaulting to '9100'".format(_mtu)
+ )
+ _mtu = 9100
+ else:
+ raise CheckerException(
+ "Negative MTU values not supported: '{}'".format(
+ _mtu
+ )
+ )
+
# init mapper
_skip, _skip_file = args_utils.get_skip_args(args)
_class = _selectClass(_env, strClassHint="mapper")
@@ -255,31 +303,38 @@
skip_list_file=_skip_file
)
# init pinger
- _pinger = pinger.NetworkPinger(
+ _class = _selectClass(_env, strClassHint="pinger")
+ _pinger = _class(
_mapper,
- mtu=args.mtu,
detailed=args.detailed,
skip_list=_skip,
skip_list_file=_skip_file
)
+ # Iterate networks to be checked
+ logger_cli.info("# Checking {} networks".format(len(_cidr)))
+ _results = {}
+ for _t_cidr in _cidr:
+ logger_cli.info("-> '{}'".format(_t_cidr))
+ _results[_t_cidr] = _pinger.ping_nodes(_t_cidr, _mtu)
- _ret = _pinger.ping_nodes(_cidr)
+ logger_cli.info("\n# Summary")
+ _dataset = []
+ for _cidr, _data in _results.items():
+ if not _data:
+ # no need to save the iterations and summary
+ _dataset.append(False)
+ else:
+ # print a report for cidr
+ if args.detailed:
+ # if set, print details
+ _pinger.print_details(_cidr, _data)
+ else:
+ _pinger.print_summary(_cidr, _data)
- if _ret < 0:
- # no need to save the iterations and summary
- return
- else:
+ if any(_dataset):
# save what was collected
_pinger.mapper.errors.save_iteration_data()
-
- # print a report
- _pinger.print_summary()
-
- # if set, print details
- if args.detailed:
- _pinger.print_details()
-
- return
+ return
def do_trace(args, config):
diff --git a/cfg_checker/modules/network/mapper.py b/cfg_checker/modules/network/mapper.py
index f0fd78d..875f633 100644
--- a/cfg_checker/modules/network/mapper.py
+++ b/cfg_checker/modules/network/mapper.py
@@ -769,6 +769,7 @@
skip_list_file=None
):
self.master = KubeNodes(config)
+ self.daemonset = None
super(KubeNetworkMapper, self).__init__(
config,
errors_class=errors_class,
@@ -776,22 +777,25 @@
skip_list_file=skip_list_file
)
+ def get_daemonset(self):
+ if not self.daemonset:
+ _d = self.master.prepare_daemonset("daemonset_template.yaml")
+
+ # wait for daemonset, normally less than 60 sec for all
+ # but still, let us give it 10 second per pod
+ _timeout = self.master.nodes.__len__() * 10
+ if not self.master.wait_for_daemonset(_d, timeout=_timeout):
+ raise KubeException("Daemonset deployment fail")
+ self.daemonset = _d
+ return self.daemonset
+
def get_script_output(self, script, args=None):
"""
Get runtime network by creating DaemonSet with Host network parameter
"""
# prepare daemonset
logger_cli.info("-> Preparing daemonset to get node info")
- _daemonset = self.master.prepare_daemonset(
- "daemonset_template.yaml",
- config_map=script
- )
-
- # wait for daemonset, normally less than 60 sec for all
- # but still, let us give it 10 second per pod
- _timeout = self.master.nodes.__len__() * 10
- if not self.master.wait_for_daemonset(_daemonset, timeout=_timeout):
- raise KubeException("Daemonset deployment fail")
+ _daemonset = self.get_daemonset()
logger_cli.info("-> Running script on daemonset")
# exec script on all pods in daemonset
_result = self.master.execute_script_on_daemon_set(
@@ -801,7 +805,8 @@
)
# delete daemonset
- self.master.delete_daemonset(_daemonset)
+ # TODO: handle daemonset delete properly
+ # self.master.delete_daemonset(_daemonset)
return _result
@@ -810,9 +815,11 @@
self.map_network(self.RUNTIME)
def map_network(self, source):
+ # if network type is mapped - just return it
+ if source in self.networks:
+ return self.networks[source]
# maps target network using given source
_networks = None
-
if source == self.RUNTIME:
logger_cli.info("# Mapping node runtime network data")
_r = self.get_script_output("ifs_data.py", args="json")
diff --git a/cfg_checker/modules/network/pinger.py b/cfg_checker/modules/network/pinger.py
index 30881da..04a5f68 100644
--- a/cfg_checker/modules/network/pinger.py
+++ b/cfg_checker/modules/network/pinger.py
@@ -10,36 +10,33 @@
def __init__(
self,
mapper,
- mtu=None,
detailed=False,
skip_list=None,
skip_list_file=None
):
logger_cli.info("# Initializing Pinger")
self.mapper = mapper
- # default MTU value
- self.target_mtu = mtu if mtu else 64
# only data
- self.packet_size = int(self.target_mtu) - 20 - 8
self.detailed_summary = detailed
+ self.results = {}
def _collect_node_addresses(self, target_net):
# use reclass model and standard methods
# to create list of nodes with target network
- _reclass = self.mapper.map_network(self.mapper.RUNTIME)
- if target_net in _reclass:
- return _reclass[target_net]
+ _networks = self.mapper.map_network(self.mapper.RUNTIME)
+ if target_net in _networks:
+ return _networks[target_net]
else:
logger_cli.info(
- "# Target network of {} not found in reclass".format(
+ "# Target network of {} not found after mapping".format(
target_net.exploded
)
)
return None
- def ping_nodes(self, network_cidr_str):
+ def _get_packets_data(self, network_cidr_str, mtu):
# Conduct actual ping using network CIDR
- logger_cli.info("# Collecting node pairs")
+ logger_cli.debug("... collecting node pairs")
_fake_if = ipaddress.IPv4Interface(str(network_cidr_str))
_net = _fake_if.network
# collect nodes and ips from reclass
@@ -75,13 +72,20 @@
if _ip not in src_ips:
if tgt_host not in _t:
_t[tgt_host] = []
+ # Handling mtu and packet size
+ if mtu == 0:
+ # Detect MTU
+ _mtu = int(tgt_if['mtu'])
+ else:
+ _mtu = mtu
+ _packet_size = int(_mtu)-20-8 if _mtu != 0 else 0
_tgt = {
"ip": _ip,
"tgt_host": tgt_host,
"ip_index": _ip_index,
"if_name": tgt_if_name,
- "mtu": self.target_mtu,
- "size": self.packet_size
+ "mtu": _mtu,
+ "size": _packet_size
}
_t[tgt_host].append(
_tgt
@@ -98,10 +102,213 @@
"check network configuration\n".format(network_cidr_str)
)
- return -1
+ return None, _packets, _nodes_total
+ else:
+ return _count, _packets, _nodes_total
+
+ def _process_result(self, src, src_data, result):
+ # Parse output
+ try:
+ _result = json.loads(result)
+ except (ValueError, TypeError):
+ self.mapper.errors.add_error(
+ self.mapper.errors.NET_NODE_NON_RESPONSIVE,
+ node=src,
+ response=result
+ )
+
+ _msg = "# ERROR: Unexpected return for '{}': '{}'\n".format(
+ src,
+ result
+ )
+
+ return False, _msg
+
+ # Handle return codes
+ for tgt_node, _tgt_ips in _result.items():
+ for _params in _tgt_ips:
+ _body = "{}({}) --{}--> {}({}@{})\n".format(
+ src,
+ src_data["if_name"],
+ _params["returncode"],
+ tgt_node,
+ _params["if_name"],
+ _params["ip"]
+ )
+ _stdout = ""
+ _stderr = ""
+ if len(_params["stdout"]) > 0:
+ _stdout = "stdout:\n{}\n".format(_params["stdout"])
+ if len(_params["stderr"]) > 0:
+ _stderr = "stderr:\n{}\n".format(_params["stderr"])
+
+ if not _params["returncode"]:
+ # 0
+ self.mapper.errors.add_error(
+ self.mapper.errors.NET_PING_SUCCESS,
+ ping_path=_body,
+ stdout=_stdout,
+ stderr=_stderr
+ )
+ elif _params["returncode"] == 68:
+ # 68 is a 'can't resove host error'
+ self.mapper.errors.add_error(
+ self.mapper.errors.NET_PING_NOT_RESOLVED,
+ ping_path=_body,
+ stdout=_stdout,
+ stderr=_stderr
+ )
+ elif _params["returncode"] > 1:
+ # >1 is when no actial (any) response
+ self.mapper.errors.add_error(
+ self.mapper.errors.NET_PING_ERROR,
+ ping_path=_body,
+ stdout=_stdout,
+ stderr=_stderr
+ )
+ else:
+ # 1 is for timeouts amd/or packet lost
+ self.mapper.errors.add_error(
+ self.mapper.errors.NET_PING_TIMEOUT,
+ ping_path=_body,
+ stdout=_stdout,
+ stderr=_stderr
+ )
+ return True, _result
+
+ def print_summary(self, cidr, data):
+ # Print summary of ping activity in node-node perspective
+ logger_cli.info("\n-> {}, {} nodes".format(cidr, len(data)))
+ # iterate nodes
+ for _n, _d in data.items():
+ # targets
+ _total = len(_d['targets'])
+ _fail = []
+ _pass = []
+ _mtus = []
+ for _f, _t in _d['targets'].items():
+ # filter data
+ _fail += [[_f, _l] for _l in _t if _l['returncode']]
+ _pass += [[_f, _l] for _l in _t if not _l['returncode']]
+ _mtus += [str(_l['mtu']) for _l in _t]
+ # summary of passed
+ # source node
+ logger_cli.info(
+ " PASS: {}/{} nodes from {} ({}:{}) with MTU {}".format(
+ len(_pass),
+ _total,
+ _n,
+ _d['if_name'],
+ _d['ip'],
+ ",".join(list(set(_mtus)))
+ )
+ )
+ # print fails
+ for node, _dict in _fail:
+ logger_cli.info(
+ "\tFAIL: {} ({}:{}) with {}/{}".format(
+ node,
+ _dict['if_name'],
+ _dict['ip'],
+ _dict['size'],
+ _dict['mtu']
+ )
+ )
+
+ # logger_cli.info(self.mapper.errors.get_summary(print_zeros=False))
+ return
+
+ def print_details(self, cidr, data):
+ def _print_stds(stdout, stderr):
+ logger_cli.debug(" stdout:\n")
+ for line in stdout.splitlines():
+ logger_cli.debug(" {}".format(line))
+ if not stderr:
+ logger_cli.debug(" stderr: <empty>")
+ else:
+ logger_cli.debug(" stderr:\n")
+ for line in stderr.splitlines():
+ logger_cli.debug(" {}".format(line))
+
+ logger_cli.info("\n---> {}, {} nodes".format(cidr, len(data)))
+ # iterate nodes
+ for _n, _d in data.items():
+ # targets
+ _fail = []
+ _pass = []
+ for _f, _t in _d['targets'].items():
+ # filter data
+ _fail += [[_f, _l] for _l in _t if _l['returncode']]
+ _pass += [[_f, _l] for _l in _t if not _l['returncode']]
+ # summary of passed
+ # source node
+ logger_cli.info(
+ "======= from {} ({}:{})".format(
+ _n,
+ _d['if_name'],
+ _d['ip']
+ )
+ )
+ for node, _dict in _pass:
+ logger_cli.info(
+ " + PASS: to {} ({}:{}) with {}/{}".format(
+ node,
+ _dict['if_name'],
+ _dict['ip'],
+ _dict['size'],
+ _dict['mtu']
+ )
+ )
+ _print_stds(_dict['stdout'], _dict['stderr'])
+ # print fails
+ for node, _dict in _fail:
+ logger_cli.info(
+ " - FAIL: to {} ({}:{}) with {}/{}".format(
+ node,
+ _dict['if_name'],
+ _dict['ip'],
+ _dict['size'],
+ _dict['mtu']
+ )
+ )
+ _print_stds(_dict['stdout'], _dict['stderr'])
+
+ # Detailed errors
+ # logger_cli.info(
+ # "\n{}\n".format(
+ # self.mapper.errors.get_errors()
+ # )
+ # )
+ return
+
+
+class SaltNetworkPinger(NetworkPinger):
+ def __init__(
+ self,
+ mapper,
+ detailed=False,
+ skip_list=None,
+ skip_list_file=None
+ ):
+ super(SaltNetworkPinger, self).__init__(
+ mapper,
+ detailed=detailed,
+ skip_list=skip_list,
+ skip_list_file=skip_list_file
+ )
+
+ def ping_nodes(self, network_cidr_str, mtu):
+ # get packets
+ _count, _packets, _nodes_total = self._get_packets_data(
+ network_cidr_str,
+ mtu
+ )
+ if not _count:
+ return None
# do ping of packets
- logger_cli.info("# Pinging nodes: MTU={}".format(self.target_mtu))
+ logger_cli.info(
+ "-> Pinging nodes in {}".format(network_cidr_str))
self.mapper.master.prepare_script_on_active_nodes("ping.py")
_progress = Progress(_count)
_progress_index = 0
@@ -131,89 +338,96 @@
src
)
)
- # Parse salt output
- _result = _results[src]
- try:
- _result = json.loads(_result)
- except (ValueError, TypeError):
- _progress.clearline()
- logger_cli.error(
- "# ERROR: Unexpected salt return for '{}': '{}'\n".format(
- src,
- _result
- )
- )
- self.mapper.errors.add_error(
- self.mapper.errors.NET_NODE_NON_RESPONSIVE,
- node=src,
- response=_result
- )
- continue
- # Handle return codes
- for tgt_node, _tgt_ips in _result.items():
- for _params in _tgt_ips:
- _body = "{}({}) --{}--> {}({}@{})\n".format(
- src,
- src_data["if_name"],
- _params["returncode"],
- tgt_node,
- _params["if_name"],
- _params["ip"]
- )
- _stdout = ""
- _stderr = ""
- if len(_params["stdout"]) > 0:
- _stdout = "stdout:\n{}\n".format(_params["stdout"])
- if len(_params["stderr"]) > 0:
- _stderr = "stderr:\n{}\n".format(_params["stderr"])
-
- if not _params["returncode"]:
- # 0
- self.mapper.errors.add_error(
- self.mapper.errors.NET_PING_SUCCESS,
- ping_path=_body,
- stdout=_stdout,
- stderr=_stderr
- )
- elif _params["returncode"] == 68:
- # 68 is a 'can't resove host error'
- self.mapper.errors.add_error(
- self.mapper.errors.NET_PING_NOT_RESOLVED,
- ping_path=_body,
- stdout=_stdout,
- stderr=_stderr
- )
- elif _params["returncode"] > 1:
- # >1 is when no actial (any) response
- self.mapper.errors.add_error(
- self.mapper.errors.NET_PING_ERROR,
- ping_path=_body,
- stdout=_stdout,
- stderr=_stderr
- )
- else:
- # 1 is for timeouts amd/or packet lost
- self.mapper.errors.add_error(
- self.mapper.errors.NET_PING_TIMEOUT,
- ping_path=_body,
- stdout=_stdout,
- stderr=_stderr
- )
-
# Parse results back in place
- src_data["targets"] = _result
+ _ret_code, _data = self._process_result(
+ src, src_data,
+ _results[src]
+ )
+ if not _ret_code:
+ _progress.clearline()
+ logger_cli.error(_data)
+ continue
+ else:
+ src_data["targets"] = _data
_progress.end()
- return 0
+ return _packets
- def print_summary(self):
- logger_cli.info(self.mapper.errors.get_summary(print_zeros=False))
- def print_details(self):
- # Detailed errors
- logger_cli.info(
- "\n{}\n".format(
- self.mapper.errors.get_errors()
- )
+class KubeNetworkPinger(NetworkPinger):
+ def __init__(
+ self,
+ mapper,
+ detailed=False,
+ skip_list=None,
+ skip_list_file=None
+ ):
+ super(KubeNetworkPinger, self).__init__(
+ mapper,
+ detailed=detailed,
+ skip_list=skip_list,
+ skip_list_file=skip_list_file
)
+
+ def ping_nodes(self, network_cidr_str, mtu):
+ # get packets
+ _count, _packets, _nodes_total = self._get_packets_data(
+ network_cidr_str,
+ mtu
+ )
+ if not _count:
+ return None
+
+ # do ping of packets
+ logger_cli.info(
+ "-> Pinging nodes in {}".format(network_cidr_str))
+ _progress = Progress(_count)
+ _progress_index = 0
+ _node_index = 0
+ for src, src_data in _packets.items():
+ _targets = src_data["targets"]
+ _node_index += 1
+ # create 'targets.json' on source host
+ _ds = self.mapper.get_daemonset()
+ _pname = self.mapper.master.get_pod_name_in_daemonset_by_node(
+ src,
+ _ds
+ )
+ _path = self.mapper.master.prepare_json_in_pod(
+ _pname,
+ _ds.metadata.namespace,
+ _targets,
+ "targets.json"
+ )
+ # execute ping.py
+ _result = self.mapper.master.exec_on_target_pod(
+ _pname,
+ "ping.py",
+ args=[_path]
+ )
+ _progress_index += len(_targets)
+ # print progress
+ _progress.write_progress(
+ _progress_index,
+ note='/ {}/{} nodes / current {}'.format(
+ _node_index,
+ _nodes_total,
+ src
+ )
+ )
+ # Parse results back in place
+ _ret_code, _data = self._process_result(
+ src, src_data,
+ _result
+ )
+ if not _ret_code:
+ _progress.clearline()
+ logger_cli.error(_data)
+ continue
+ else:
+ src_data["targets"] = _data
+
+ _progress.end()
+
+ return _packets