def check_version(self, filename, branch, branch_base, arg_var, found_version, branch_version):
self.branches_dockerfile_checked.add(branch)
self.dockerfiles_checked.add(filename)
if arg_var:
log.debug("found arg '%s'", arg_var)
arg_version = "ARG '{0}={1}'".format(arg_var, found_version)
else:
arg_version = "'{0}'".format(found_version)
#log.debug("arg '%s' matches branch base '%s'", argversion.group(1), branch_base)
log.debug("comparing '%s' contents to version derived from branch '%s' => '%s'",
filename, branch, branch_version)
if not isVersion(branch_version.lstrip('jdk').lstrip('jre')):
die("unrecognized branch version '{0}' for branch_base '{1}'"
.format(branch_version, branch_base))
#if branch_version == found_version or branch_version == found_version.split('.', 2)[0]:
if found_version[0:len(branch_version)] == branch_version:
log.info("{0} version '{1}' matches {2}".
format(self.valid_git_branches_msg, branch_version, arg_version))
else:
log.error("{0} version '{1}' vs Dockerfile {2}".
format(self.invalid_git_branches_msg, branch_version, arg_version))
self.dockerfiles_failed += 1
self.branches_failed.add(branch)
return False
return True
def __parse_args__(self):
try:
(self.options, self.args) = self.__parser.parse_args()
# I don't agree with zero exit code from OptionParser for help/usage,
# and want UNKNOWN not CRITICAL(2) for switch mis-usage...
except SystemExit: # pragma: no cover
sys.exit(ERRORS['UNKNOWN'])
if self.options.help: # pragma: no cover
self.usage()
if self.options.version: # pragma: no cover
print('%(version)s' % self.__dict__)
sys.exit(ERRORS['UNKNOWN'])
if 'timeout' in dir(self.options):
self.timeout = self.get_opt('timeout')
env_verbose = os.getenv('VERBOSE')
if isInt(env_verbose):
if env_verbose > self.verbose:
log.debug('environment variable $VERBOSE = %s, increasing verbosity', env_verbose)
self.verbose = env_verbose
elif env_verbose is None:
pass
else:
log.warning("$VERBOSE environment variable is not an integer ('%s')", env_verbose)
self.parse_args()
return self.options, self.args
def consume(self):
self.check_connection()
self.check_channel()
def connection_timeout_handler():
raise CriticalError("unique message not returned on queue '{queue}' within {secs:.2f} secs"\
.format(queue=self.queue, secs=self.timeout / 3) + \
", consumer timed out while consuming messages from {name} broker '{host}:{port}'"\
.format(name=self.name, host=self.host, port=self.port))
self.conn.add_timeout(self.timeout / 3, connection_timeout_handler)
# don't re-declare, queue should still exist otherwise error out
#channel.queue_declare(queue = 'hello')
# don't ack as messages could stay in queue indefinitely
self.consumer_tag = self.channel.basic_consume(self.consumer_callback,
queue=self.queue,
# let broker autogenerate consumer_tag
# consumer_tag = self.consumer_tag),
no_ack=self.no_ack
)
# could also use non-callback mechanism - generator that yields tuples (method, properties, body)
# requires self.channel.cancel() from within loop
# self.channel.consume(self.queue,
# no_ack = True,
# exclusive = True,
# arguments = None,
# inactivity_timeout = self.timeout/3)
log.debug('start consuming')
self.channel.start_consuming()
# could instead use basic_get to return single message
# self.channel.basic_get(queue = self.queue, no_ack = True)
log.info('closing connection to broker')
self.conn.close(reply_code=200, reply_text='Normal shutdown')
return self.consumed_message
def check_http(self, host, port, url_path=''):
if not isStr(url_path):
url_path = ''
url = '{protocol}://{host}:{port}/{url_path}'.format(protocol=self.protocol,
host=host,
port=port,
url_path=url_path.lstrip('/'))
log.info('GET %s', url)
try:
# timeout here isn't total timeout, it's response time
req = requests.get(url, timeout=self.request_timeout)
except requests.exceptions.RequestException:
return False
except IOError:
return False
log.debug("%s - response: %s %s", url, req.status_code, req.reason)
log.debug("%s - content:\n%s\n%s\n%s", url, '='*80, req.content.strip(), '='*80)
if req.status_code != 200:
return None
if self.regex:
log.info('%s - checking regex against content', url)
if self.regex.search(req.content):
log.info('%s - regex matched http output', url)
else:
log.info('%s - regex did not match http output', url)
return None
log.info("%s - passed all checks", url)
return (host, port)
def check_git_tags_dockerfiles(self, target):
target = os.path.abspath(target)
gitroot = find_git_root(target)
log.debug("finding tags for target '{0}'".format(target))
repo = git.Repo(gitroot)
tags = [str(x).split("/")[-1] for x in repo.tags]
if self.tag_prefix is not None:
log.debug("restricting to tags matching tag prefix")
tags = [x for x in tags if self.tag_prefix.match(x)]
# if log.isEnabledFor(logging.DEBUG):
log.debug("\n\ntags for target %s:\n\n%s\n", target, "\n".join(tags))
original_checkout = "master"
try:
try:
original_checkout = repo.active_branch.name
except TypeError as _:
pass
for tag in tags:
log.debug("checking tag '%s' Dockerfiles for target '%s'", tag, target)
try:
repo.git.checkout(tag)
except git.exc.GitCommandError as _:
die(_)
self.check_path(target, tag)
except Exception as _: # pylint: disable=broad-except
die(_)
finally:
log.debug("returning to original checkout '%s'", original_checkout)
repo.git.checkout(original_checkout)
def __init__(self, arg, **kwargs):
# min=None, max=None, positive=True, integer=True, simple='upper', name='', **kwargs):
self.name = kwargs.get("name", "")
if self.name:
self.name += " "
self.thresholds = {"upper": None, "lower": None}
self.opts = {"invert": False}
self.opts["simple"] = kwargs.get("simple", "upper")
self.opts["positive"] = kwargs.get("positive", True)
self.opts["integer"] = kwargs.get("integer", True)
self.thresholds["min"] = kwargs.get("min", None)
self.thresholds["max"] = kwargs.get("max", None)
log.debug("%sthreshold simple = %s", self.name, self.opts["simple"])
log.debug("%sthreshold positive = %s", self.name, self.opts["positive"])
log.debug("%sthreshold integer = %s", self.name, self.opts["integer"])
log.debug("%sthreshold min = %s", self.name, self.thresholds["min"])
log.debug("%sthreshold max = %s", self.name, self.thresholds["max"])
if self.opts["simple"] not in ("upper", "lower"):
raise CodingError("simple threshold type must be one of: upper, lower")
if not isBool(self.opts["positive"]):
raise CodingError("positive option must be set to either True or False")
if not isBool(self.opts["integer"]):
raise CodingError("integer option must be set to either True or False")
self.__parse_threshold__(arg, kwargs.get("optional"))
def check_path(self, path, tag):
status = True
if os.path.isfile(path):
return self.check_file(path, tag)
elif os.path.isdir(path):
if os.path.basename(path) == ".git":
return True
for item in os.listdir(path):
subpath = os.path.join(path, item)
if os.path.islink(subpath):
subpath = os.path.realpath(subpath)
if os.path.isdir(subpath):
tag_base = tag.rsplit("-", 1)[0]
subpath_base = os.path.basename(subpath)
# log.debug('tag_base = %s', tag_base)
# log.debug('subpath_base = %s', subpath_base)
if subpath_base == tag_base:
if not self.check_path(subpath, tag):
status = False
elif os.path.isfile(subpath):
if not self.check_file(subpath, tag):
status = False
elif not os.path.exists(subpath):
log.debug("subpath '%s' does not exist in tag '%s', skipping..." % (subpath, tag))
else:
die("failed to determine if subpath '%s' is file or directory in tag '%s'" % (subpath, tag))
elif not os.path.exists(path):
log.debug("path '%s' does not exist in tag '%s', skipping..." % (path, tag))
else:
die("failed to determine if path '%s' is file or directory in tag '%s'" % (path, tag))
return status
def parse(self, stdout):
output = [_ for _ in stdout.split('\n') if _]
if len(output) < 2:
raise CriticalError("docker image '{repo}' not found! Does not exist or has not been pulled yet?"\
.format(repo=self.docker_image))
name_len = len(self.docker_image)
if len(output) > 2:
tags = set([line[name_len:name_len + 10].strip() for line in output[1:]])
tags = [tag for tag in tags if tag != '<none>']
tags = sorted(list(tags))
if log.isEnabledFor(logging.DEBUG):
for tag in tags:
log.debug('found tag: %s', tag)
raise UnknownError('too many results returned - did you forget to suffix a specific :tag to ' + \
'--docker-image? (eg. :latest, :1.1). The following tags were found: ' + \
', '.join(tags)
)
header_line = output[0]
docker_image_line = output[1]
image_header = ' '.join(header_line.split()[2:4])
log.debug('image header column: %s', image_header)
if image_header != 'IMAGE ID':
raise UnknownError("3rd column in header '{0}' is not 'IMAGE ID' as expected, parsing failed!"\
.format(image_header))
self.msg = "docker image '{repo}'".format(repo=self.docker_image)
self.check_id(docker_image_line)
self.check_size(docker_image_line)
def check_file(self, filename, tag):
filename = os.path.abspath(filename)
if os.path.basename(filename) != "Dockerfile":
return True
parent = os.path.basename(os.path.dirname(filename))
tag_base = tag.rsplit("-", 1)[0]
if parent.lower() != tag_base.lower():
log.debug(
"skipping '{0}' as it's parent directory '{1}' doesn't match tag base '{2}'".format(
filename, parent, tag_base
)
)
return True
self.valid_git_tags_msg = "%s => Dockerfile Git Tags OK" % filename
self.invalid_git_tags_msg = "%s => Dockerfile Git Tags MISMATCH in tag '%s'" % (filename, tag)
try:
if not self.check_dockerfile_arg(filename, tag):
self.failed = True
# print(self.invalid_git_tags_msg)
return False
# now switched to per tag scan this returns way too much redundant output
# print(self.valid_git_tags_msg)
except IOError as _:
die("ERROR: %s" % _)
return True
def check_media_file(self, filename):
valid_media_msg = '%s => OK' % filename
invalid_media_msg = '%s => INVALID' % filename
try:
# cmd = self.validate_cmd.format(filename)
cmd = self.validate_cmd
log.debug('cmd: %s %s', cmd, filename)
log.info('verifying {0}'.format(filename))
# capturing stderr to stdout because ffprobe prints to stderr in all cases
# Python 2.7+
#subprocess.check_output(cmd.split() + [filename], stderr=subprocess.STDOUT)
proc = subprocess.Popen(cmd.split() + [filename], stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
(stdout, _) = proc.communicate()
returncode = proc.wait()
if returncode != 0 or (stdout is not None and 'Error' in stdout):
_ = CalledProcessError(returncode, cmd)
_.output = stdout
raise _
print(valid_media_msg)
except CalledProcessError as _:
if self.verbose > 2:
print(_.output)
if self.skip_errors:
print(invalid_media_msg)
self.failed = True
return False
die(invalid_media_msg)
def check_multirecord_json(self):
log.debug('check_multirecord_json()')
for line in self.iostream:
if isJson(line):
# can't use self.print() here, don't want to print valid for every line of a file / stdin
if self.passthru:
print(line, end='')
elif isJson(line.replace("'", '"')):
if self.permit_single_quotes:
log.debug('valid multirecord json (single quoted)')
# self.single_quotes_detected = True
if self.passthru:
print(line, end='')
else:
log.debug('invalid multirecord json (single quoted)')
self.failed = True
if not self.passthru:
die('%s (multi-record format)' % self.invalid_json_msg_single_quotes)
return False
else:
log.debug('invalid multirecord json')
self.failed = True
return False
# self.multi_record_detected = True
log.debug('multirecord json (all lines passed)')
if not self.passthru:
print('%s (multi-record format)' % self.valid_json_msg)
return True
def check_git_branches_upstream(self, target):
target = os.path.abspath(target)
gitroot = find_git_root(target)
if gitroot is None:
die('Failed to find git root for target {0}'.format(target))
log.debug("finding branches for target '{0}'".format(target))
repo = git.Repo(gitroot)
branches = repo.branches
if self.branch_prefix is not None:
log.debug('restricting to branches matching branch prefix')
branches = [x for x in branches if self.branch_prefix.match(str(x))]
if not branches:
log.error("No branches matching '%s' for target '%s'", self.get_opt('branch_prefix'), target)
self.status = 'NO BRANCHES'
#if log.isEnabledFor(logging.DEBUG):
#log.debug('\n\nbranches for target %s:\n\n%s\n', target, '\n'.join(list(branches)))
for branch in branches:
expected = '{0}/{1}'.format(self.origin, branch)
tracking_branch = str(branch.tracking_branch())
if tracking_branch == expected:
log.info("OK: repo '{0}' branch '{1}' is tracking '{2}'"
.format(gitroot, branch, tracking_branch))
else:
self.status = "ERROR"
log.error("BAD: branch '{0}' is tracking '{1}' (expected '{2}')"
.format(branch, tracking_branch, expected))
def check(self, client):
log.info('running Docker info')
info = client.info()
if log.isEnabledFor(logging.DEBUG):
log.debug(jsonpp(info))
containers = info['Containers']
running_containers = info['ContainersRunning']
paused_containers = info['ContainersPaused']
stopped_containers = info['ContainersStopped']
self.msg = 'Docker '
if self.running:
self.msg += 'running containers = {}'.format(running_containers)
self.check_thresholds(running_containers)
self.msg += ' | running_containers={}{}'.format(running_containers, self.get_perf_thresholds())
elif self.paused:
self.msg += 'paused containers = {}'.format(paused_containers)
self.check_thresholds(paused_containers)
self.msg += ' | paused_containers={}{}'.format(paused_containers, self.get_perf_thresholds())
elif self.stopped:
self.msg += 'stopped containers = {}'.format(stopped_containers)
self.check_thresholds(stopped_containers)
self.msg += ' | stopped_containers={}{}'.format(stopped_containers, self.get_perf_thresholds())
elif self.total:
self.msg += 'total containers = {}'.format(containers)
self.check_thresholds(containers)
self.msg += ' | total_containers={}{}'.format(containers, self.get_perf_thresholds())
else:
self.msg += 'containers = {}, running containers = {}, paused containers = {}, stopped containers = {}'\
.format(containers, running_containers, paused_containers, stopped_containers)
self.msg += ' | containers={} running_containers={} paused_containers={} stopped_containers={}'\
.format(containers, running_containers, paused_containers, stopped_containers)
def check_path(self, path, branch):
status = True
(branch_base, _) = self.branch_version(branch)
if os.path.isfile(path):
return self.check_file(path, branch)
elif os.path.isdir(path):
if os.path.basename(path) == '.git':
return True
for item in os.listdir(path):
subpath = os.path.join(path, item)
if os.path.islink(subpath):
subpath = os.path.realpath(subpath)
if os.path.isdir(subpath):
subpath_base = os.path.basename(subpath)
#log.debug('subpath_base = %s', subpath_base)
if self.normalize_name(subpath_base) == self.normalize_name(branch_base):
if not self.check_path(subpath, branch):
status = False
elif os.path.isfile(subpath):
if not self.check_file(subpath, branch):
status = False
elif not os.path.exists(subpath):
log.debug("subpath '%s' does not exist in branch '%s', skipping..." % (subpath, branch))
else:
die("failed to determine if subpath '%s' is file or directory in branch '%s'" % (subpath, branch))
elif not os.path.exists(path):
log.debug("path '%s' does not exist in branch '%s', skipping..." % (path, branch))
else:
die("failed to determine if path '%s' is file or directory in branch '%s'" % (path, branch))
return status
def timeout_max(self, secs):
if secs is not None and not isInt(secs):
raise CodingError('invalid timeout max passed to set_timeout_max(), must be an integer representing seconds') # pylint: disable=line-too-long
# leave this to be able to set max to any amount
# validate_int(secs, 'timeout default', 0, self.__timeout_max )
log.debug('setting max timeout to %s secs', secs)
self.__timeout_max = secs
def subscribe(self):
self.consumer = KafkaConsumer(
#self.topic,
bootstrap_servers=self.brokers,
# client_id=self.client_id,
# group_id=self.group_id,
request_timeout_ms=self.timeout_ms
)
#key_serializer
#value_serializer
log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
# log.debug('subscribing to topic \'{0}\' parition \'{1}\''.format(self.topic, self.partition))
# self.consumer.subscribe(TopicPartition(self.topic, self.partition))
# log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
log.debug('assigning partition {0} to consumer'.format(self.partition))
# self.consumer.assign([self.partition])
self.consumer.assign([self.topic_partition])
log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
log.debug('getting current offset')
self.start_offset = self.consumer.position(self.topic_partition)
# self.start_offset = 0
if self.start_offset is None:
self.start_offset = 0
#raise UnknownError('Kafka Consumer reported current starting offset = {0}'.format(self.start_offset))
log.debug('recorded starting offset \'{0}\''.format(self.start_offset))
def check_git_branches_dockerfiles(self, target):
gitroot = find_git_root(target)
if gitroot is None:
die('Failed to find git root for target {0}'.format(target))
log.debug("finding branches for target '{0}'".format(target))
repo = git.Repo(gitroot)
branches = [str(x) for x in repo.refs if isinstance(x, git.refs.remote.RemoteReference)]
branches = [x.split('/')[-1] for x in branches]
branches = [x for x in branches if x not in ('HEAD', 'master')]
if self.branch_prefix is not None:
log.debug('restricting to branches matching branch prefix')
branches = [x for x in branches if self.branch_prefix.match(x)]
#if log.isEnabledFor(logging.DEBUG):
log.debug('\n\nbranches for target %s:\n\n%s\n', target, '\n'.join(branches))
original_checkout = 'master'
try:
try:
original_checkout = repo.active_branch.name
except TypeError as _:
pass
for branch in branches:
log.debug("checking branch '%s' Dockerfiles for target '%s'", branch, target)
self.branches_checked += 1
try:
repo.git.checkout(branch)
except git.exc.GitCommandError as _:
die(_)
self.check_path(target, branch)
except Exception as _: # pylint: disable=broad-except
traceback.print_exc()
sys.exit(1)
finally:
log.debug("returning to original checkout '%s'", original_checkout)
repo.git.checkout(original_checkout)
def run(self):
server_url = '{proto}://{host}:{port}'.format(proto=self.protocol, host=self.host, port=self.port)
try:
log.debug('setting up Jenkins connection to %s', server_url)
start_time = time.time()
server = jenkins.Jenkins(server_url, username=self.user, password=self.password, timeout=self.timeout / 3)
if log.isEnabledFor(logging.DEBUG):
log.debug('getting user')
user = server.get_whoami()
log.debug('connected as user %s', jsonpp(user))
log.debug('getting plugin info')
#plugins = server.get_plugins()
# deprecated but .get_plugins() output is not JSON serializable
# so must use old deprecated method get_plugins_info() :-/
plugins = server.get_plugins_info()
query_time = time.time() - start_time
except jenkins.JenkinsException as _:
raise CriticalError(_)
if log.isEnabledFor(logging.DEBUG):
log.debug('%s', jsonpp(plugins))
plugin_count = len(plugins)
update_count = 0
for plugin in plugins:
if plugin['hasUpdate']:
update_count += 1
self.msg += " {0} plugin update{1} available out of {2} installed plugin{3}".format(update_count,
plural(update_count),
plugin_count,
plural(plugin_count))
if update_count:
self.warning()
self.msg += ' | updates_available={0};1 plugins_installed={1} query_time={2:.4f}s'.format(update_count,
plugin_count,
query_time)
def main(self):
try:
# Python 2.x
super(NagiosPlugin, self).main()
# Python 3.x
# super().__init__()
# redirect_stderr_stdout()
except CriticalError as _:
qquit('CRITICAL', _)
except WarningError as _:
qquit('WARNING', _)
except UnknownError as _:
qquit('UNKNOWN', _)
except CodingError as _:
qquit('UNKNOWN', 'Programming Error: {0}. {1}'.format(_, support_msg()))
except Exception as _: # pylint: disable=broad-except
exception_type = type(_).__name__
if log.isEnabledFor(logging.DEBUG):
log.debug("exception: '%s'", exception_type)
log.debug(traceback.format_exc())
msg = 'Nagios Plugin Exception: {exception_type}: {msg}'.format(exception_type=exception_type, msg=self.exception_msg())
#msg = ', '.join([x.strip() for x in msg.split('\n')])
# ', ' doesn't look nice for ':\n ...' => ':, ...' (snakebite OutOfNNException)
#msg = '\t'.join([x.strip() for x in msg.split('\n')])
#if self.options.verbose > 2:
# msg = type(_).__name__ + ': ' + msg
msg += '. ' + support_msg()
qquit('UNKNOWN', msg)
def subscribe(self):
self.consumer = KafkaConsumer(
#self.topic,
bootstrap_servers=self.brokers,
# client_id=self.client_id,
# group_id=self.group_id,
request_timeout_ms=self.timeout_ms
)
#key_serializer
#value_serializer
# this is only a guess as Kafka doesn't expose it's API version
#log.debug('kafka api version: %s', self.consumer.config['api_version'])
log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
# log.debug('subscribing to topic \'{0}\' parition \'{1}\''.format(self.topic, self.partition))
# self.consumer.subscribe(TopicPartition(self.topic, self.partition))
# log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
log.debug('assigning partition {0} to consumer'.format(self.partition))
# self.consumer.assign([self.partition])
self.consumer.assign([self.topic_partition])
log.debug('partition assignments: {0}'.format(self.consumer.assignment()))
log.debug('getting current offset')
# see also highwater, committed, seek_to_end
self.start_offset = self.consumer.position(self.topic_partition)
if self.start_offset is None:
# don't do this, I've seen scenario where None is returned and all messages are read again, better to fail
# log.warn('consumer position returned None, resetting to zero')
# self.start_offset = 0
raise UnknownError('Kafka Consumer reported current starting offset = {0}'.format(self.start_offset))
log.debug('recorded starting offset \'{0}\''.format(self.start_offset))
def run(self):
server_url = '{proto}://{host}:{port}'.format(proto=self.protocol, host=self.host, port=self.port)
try:
log.debug('setting up Jenkins connection to %s', server_url)
start_time = time.time()
server = jenkins.Jenkins(server_url, username=self.user, password=self.password, timeout=self.timeout / 3)
if log.isEnabledFor(logging.DEBUG):
log.debug('getting user')
user = server.get_whoami()
log.debug('connected as user %s', jsonpp(user))
if self.list_jobs:
log.debug('getting jobs')
#jobs = server.get_jobs()
# recursively get all jobs
jobs = server.get_all_jobs()
# more efficient with many folders
# jobs = server.run_script("""
# import groovy.json.JsonBuilder;
#
# // get all projects excluding matrix configuration
# // as they are simply part of a matrix project.
# // there may be better ways to get just jobs
# items = Jenkins.instance.getAllItems(AbstractProject);
# items.removeAll {
# it instanceof hudson.matrix.MatrixConfiguration
# };
#
# def json = new JsonBuilder()
# def root = json {
# jobs items.collect {
# [
# name: it.name,
# url: Jenkins.instance.getRootUrl() + it.getUrl(),
# color: it.getIconColor().toString(),
# fullname: it.getFullName()
# ]
# }
# }
#
# // use json.toPrettyString() if viewing
# println json.toString()
# """)
print('Jenkins Jobs:\n')
for job in jobs:
print(job['fullname'])
sys.exit(ERRORS['UNKNOWN'])
log.debug('checking job exists')
if server.job_exists(self.job):
self.msg += 'exists'
else:
self.critical()
self.msg += 'does not exist!'
except jenkins.JenkinsException as _:
raise CriticalError(_)
query_time = time.time() - start_time
self.msg += ' | query_time={0:.4f}s'.format(query_time)
def process_json(self, content):
try:
self.json_data = json.loads(content)
if log.isEnabledFor(logging.DEBUG):
log.debug('JSON prettified:\n\n%s\n%s', jsonpp(self.json_data), '='*80)
return self.parse_json(self.json_data)
except (KeyError, ValueError) as _:
#raise UnknownError('{0}: {1}. {2}'.format(type(_).__name__, _, support_msg_api()))
raise UnknownError('{0}. {1}'.format(self.exception_msg(), support_msg_api()))
def parse_table(table):
for row in table.findChildren('tr'):
for col in row.findChildren('td'):
if 'Regions in Transition for more than ' in col.get_text():
log.debug('found Regions in Transition for more than ... getting next td')
next_sibling = col.findNext('td')
regions_stuck_in_transition = next_sibling.get_text().strip()
return regions_stuck_in_transition
return None
def normalize_name(name):
# allow all -dev dirs to match same branch
# special case for solr -> solrcloud dirs
name2 = name
name2 = re.sub(pattern=r'-dev$', repl='', string=name2)
name2 = re.sub(pattern=r'cloud$', repl='', string=name2)
name2 = name2.lower()
log.debug("normalized name '%s' => '%s'", name, name2)
return name2
def check(self, client):
log.info('running Docker info')
info = client.info()
if log.isEnabledFor(logging.DEBUG):
log.debug(jsonpp(info))
images = info['Images']
self.msg = 'Docker images = {}'.format(images)
self.check_thresholds(images)
self.msg += ' | docker_images={}{}'.format(images, self.get_perf_thresholds())
def main(self):
# log.debug('running main()')
log.setLevel(logging.WARN)
self.setup()
try:
self.add_options()
self.add_default_opts()
except InvalidOptionException as _:
self.usage(_)
try:
self.__parse_args__()
# broken
# autoflush()
# too late
# os.environ['PYTHONUNBUFFERED'] = "anything"
self.verbose = self.get_opt('verbose')
if self.is_option_defined('quiet') and self.get_opt('quiet'):
self.verbose = 0
elif self.verbose > 2:
log.setLevel(logging.DEBUG)
elif self.verbose > 1:
log.setLevel(logging.INFO)
elif self.verbose > 0 and self._prog[0:6] != 'check_':
log.setLevel(logging.WARN)
if self.options.debug:
log.setLevel(logging.DEBUG) # pragma: no cover
log.debug('enabling debug logging')
if self.verbose < 3:
self.verbose = 3
log.info('Hari Sekhon %s', self.version)
log.info(self._github_repo)
log.info('verbose level: %s (%s)', self.verbose, logging.getLevelName(log.getEffectiveLevel()))
if self.timeout is not None:
validate_int(self.timeout, 'timeout', 0, self.timeout_max)
log.debug('setting timeout alarm (%s)', self.timeout)
signal.signal(signal.SIGALRM, self.timeout_handler)
signal.alarm(int(self.timeout))
# if self.options.version:
# print(self.version)
# sys.exit(ERRORS['UNKNOWN'])
self.process_options()
self.process_args()
try:
self.run()
except CriticalError as _:
qquit('CRITICAL', _)
except WarningError as _:
qquit('WARNING', _)
except UnknownError as _:
qquit('UNKNOWN', _)
self.__end__()
except InvalidOptionException as _:
self.usage(_) # pragma: no cover
except KeyboardInterrupt:
# log.debug('Caught control-c...')
print('Caught control-c...') # pragma: no cover
def get_mounts():
try:
with open('/proc/mounts', 'r') as _:
lines = _.readlines()
if log.isEnabledFor(logging.DEBUG):
for line in lines:
log.debug('/proc/mounts: %s', line.rstrip('\n'))
return lines
except IOError as _:
raise UnknownError(_)
def process_file(self, filepath):
log.debug('processing filepath \'%s\'', filepath)
if filepath == '-':
filepath = '<STDIN>'
if filepath == '<STDIN>':
self.json_to_xml(sys.stdin.read())
else:
with open(filepath) as _:
content = _.read()
print(self.json_to_xml(content, filepath=filepath))
def save_cluster(self, cluster, path=''):
# log.debug('save_cluster(%s, %s)' % (cluster, name))
if not path:
path = os.path.normpath(os.path.join(self.blueprint_dir, cluster))
data = self.get_cluster_blueprint(cluster)
# logged in save()
# log.info("saving cluster '%s' blueprint to file '%s'" % (cluster, path))
if log.isEnabledFor(logging.DEBUG):
log.debug("cluster '%s' blueprint content = '%s'" % (cluster, data))
self.save(cluster, path, data)
def save_blueprint(self, blueprint, path=''):
# log.debug('save_blueprint(%s, %s' % (blueprint, name))
if not path:
path = os.path.normpath(os.path.join(self.blueprint_dir, blueprint))
data = self.get_blueprint(blueprint)
# logged in save()
# log.info("saving blueprint '%s' to file '%s" % (blueprint, path))
if log.isEnabledFor(logging.DEBUG):
log.debug("blueprint '%s' content = '%s'" % (blueprint, data))
self.save(blueprint, path, data)
def get_failing_job_id_from_build(self, build):
log.info('getting failed job id for build %s', build['id'])
if 'jobs' not in build:
raise UnknownError(
'no jobs field found in build, {0}'.format(support_msg_api))
for _ in build['jobs']:
_id = _['id']
url = 'https://api.travis-ci.org/jobs/{id}'.format(id=_id)
req = self.request_handler.get(url)
# if this raises ValueError it'll be caught by run handler
job = json.loads(req.content)
if log.isEnabledFor(logging.DEBUG):
log.debug("job id %s status:\n%s", _id, jsonpp(job))
if job['state'] == 'finished' and job['status'] in (None, '1'):
return _id
raise UnknownError('no failed job found in build {0}'.format(
build['id']))
def run(self):
url = 'https://api.travis-ci.org/repos/{repo}/builds'.format(
repo=self.repo)
request_handler = RequestHandler()
req = request_handler.get(url)
if log.isEnabledFor(logging.DEBUG):
log.debug("\n%s", jsonpp(req.content))
try:
self.parse_results(req.content)
except (KeyError, ValueError):
exception = traceback.format_exc().split('\n')[-2]
# this covers up the traceback info and makes it harder to debug
#raise UnknownError('failed to parse expected json response from Travis CI API: {0}'.format(exception))
qquit(
'UNKNOWN',
'failed to parse expected json response from Travis CI API: {0}. {1}'
.format(exception, support_msg_api()))
def is_file_dup_by_regex(self, filepath):
#match = re.search(self.regex, filepath)
basename = os.path.basename(filepath)
match = re.search(self.regex, basename)
if match:
log.debug("regex matched file '%s'", filepath)
if match.groups():
capture = match.group(1)
else:
capture = match.group(0)
if capture in self.regex_captures:
self.dups_by_regex[capture] = self.dups_by_regex.get(capture, set())
self.dups_by_regex[capture].add(self.regex_captures[capture])
self.dups_by_regex[capture].add(filepath)
return True
self.regex_captures[capture] = filepath
return False
def check_table_regions(self):
log.info('checking regions for table \'%s\'', self.table)
regions = None
try:
table = self.conn.table(self.table)
regions = table.regions()
except HBaseIOError as _:
#if 'org.apache.hadoop.hbase.TableNotFoundException' in _.message:
if 'TableNotFoundException' in _.message:
qquit('CRITICAL', 'table \'{0}\' does not exist'.format(self.table))
else:
qquit('CRITICAL', _)
except (socket.timeout, ThriftException) as _:
qquit('CRITICAL', _)
if log.isEnabledFor(logging.DEBUG):
log.debug('%s', jsonpp(regions))
if not regions:
qquit('CRITICAL', 'failed to get regions for table \'{0}\''.format(self.table))
if not isList(regions):
qquit('UNKNOWN', 'region info returned is not a list! ' + support_msg_api())
num_regions = len(regions)
log.info('num regions: %s', num_regions)
self.msg = 'HBase table \'{0}\' has {1} region{2}'.format(self.table, num_regions, plural(num_regions))
self.check_thresholds(num_regions)
num_unassigned_regions = 0
for region in regions:
try:
if not region['server_name']:
#log.debug('region \'%s\' is not assigned to any server', region['name'])
num_unassigned_regions += 1
except KeyError as _:
qquit('UNKNOWN', 'failed to find server assigned to region. ' + support_msg_api())
log.info('num unassigned regions: %s', num_unassigned_regions)
self.msg += ', {0} unassigned region{1}'.format(num_unassigned_regions, plural(num_unassigned_regions))
if num_unassigned_regions > 0:
self.warning()
self.msg += '!'
self.msg += ' |'
self.msg += ' num_regions={0}'.format(num_regions) + self.get_perf_thresholds(boundary='lower')
self.msg += ' num_unassigned_regions={0};1;0'.format(num_unassigned_regions)
log.info('finished, closing connection')
self.conn.close()
def process_options(self):
self.no_args()
self.host = self.get_opt('host')
self.port = self.get_opt('port')
validate_host(self.host)
validate_port(self.port)
if self.auth and self.get_opt('kerberos'):
self.auth = 'kerberos'
if self.auth:
self.user = self.get_opt('user')
self.password = self.get_opt('password')
if self.auth == 'optional':
if self.user and self.password:
validate_user(self.user)
validate_password(self.password)
elif self.auth == 'kerberos':
if os.getenv('KRB5_CLIENT_KTNAME'):
log.debug('kerberos enabled, will try to use keytab at %s',
os.getenv('KRB5_CLIENT_KTNAME'))
# if using KRB5_CLIENT_KTNAME to kinit avoid clobbering the same TGT cache /tmp/krb5cc_{uid}
# as that may be used by different programs kinit'd different keytabs
os.environ[
'KRB5CCNAME'] = '/tmp/krb5cc_{euid}_{basename}'.format(
euid=os.geteuid(), basename=prog)
else:
validate_user(self.user)
validate_password(self.password)
ssl_noverify = self.get_opt('ssl_noverify')
if ssl_noverify:
log_option('ssl no verify', 'true')
ssl = 1
os.environ['SSL_NO_VERIFY'] = '1'
# doesn't work, probably too late after instantiation
#if not os.getenv('PYTHONWARNINGS'):
# os.environ['PYTHONWARNINGS'] = 'ignore:Unverified HTTPS request'
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)
else:
ssl = self.get_opt('ssl')
log_option('ssl', ssl)
if ssl and self.protocol == 'http':
self.protocol = 'https'
if self.json:
# recommended for many systems like CouchDB
# but breaks Ambari API calls
#self.headers['Accept'] = 'application/json'
self.headers['Content-Type'] = 'application/json'
def run(self):
self.no_args()
host = self.get_opt('host')
port = self.get_opt('port')
slave = self.get_opt('slave')
list_slaves = self.get_opt('list_slaves')
validate_host(host)
validate_port(port)
if not list_slaves:
validate_host(slave, 'slave')
url = 'http://%(host)s:%(port)s/master/slaves' % locals()
log.debug('GET %s', url)
try:
req = requests.get(url)
except requests.exceptions.RequestException as _:
qquit('CRITICAL', _)
log.debug("response: %s %s", req.status_code, req.reason)
log.debug("content:\n{0}\n{1}\n{2}".format('=' * 80,
req.content.strip(),
'=' * 80))
if req.status_code != 200:
qquit('CRITICAL',
"Non-200 response! %s %s" % (req.status_code, req.reason))
content = req.content
if not isJson(content):
qquit('UNKNOWN', 'invalid JSON returned by Mesos Master')
data = json.loads(content)
if log.isEnabledFor(logging.DEBUG):
log.debug('\n%s', jsonpp(data))
slaves = {}
regex = re.compile(r'^slave\(\d+\)\@(.+):\d+')
try:
for item in data['slaves']:
match = regex.match(item['pid'])
if match:
slaves[item['hostname']] = match.group(1)
else:
slaves[item['hostname']] = item['pid']
except KeyError:
qquit(
'UNKNOWN',
'failed to parse slaves from Mesos API output. {0}'.format(
support_msg_api))
if list_slaves:
qquit('UNKNOWN', 'Slaves list:\n\n{0}'.format(dict_lines(slaves)))
log.info('found slaves:\n\n{0}\n'.format(dict_lines(slaves)))
slave = slave.lower()
for _ in slaves:
if slave == _.lower() or slave == slaves[_].lower():
qquit('OK',
"Mesos slave '{0}' registered with master".format(slave))
break
else:
qquit('CRITICAL',
"Mesos slave '{0}' not registered with master".format(slave))
def process_csv(self, filehandle):
csvreader = None
try:
if self.delimiter is not None:
try:
csvreader = csv.reader(filehandle, delimiter=self.delimiter, quotechar=self.quotechar)
except TypeError as _:
self.usage(_)
else:
# dialect = csv.excel
dialect = csv.Sniffer().sniff(filehandle.read(1024))
# this will raise an Error if invalid
dialect.strict = True
filehandle.seek(0)
csvreader = csv.reader(filehandle, dialect)
except csv.Error as _:
log.warning('file %s: %s', self.filename, _)
return False
count = 0
try:
# csvreader doesn't seem to generate any errors ever :-(
# csv module allows entire lines of json/xml/yaml to go in as a single field
# Adding some invalidations manually
for field_list in csvreader:
# list of fields with no separator information
# log.debug("line: %s", _)
# make it fail if there is only a single field on any line
if len(field_list) < 2:
return False
# it's letting JSON through :-/
if field_list[0] == '{':
return False
# extra protection along the same lines as anti-json:
# the first char of field should be alphanumeric, not syntax
# however instead of isAlnum allow quotes for quoted CSVs to pass validation
if not isChars(field_list[0][0], 'A-Za-z0-9\'"'):
return False
count += 1
except csv.Error as _:
log.warning('file %s, line %s: %s', self.filename, csvreader.line_num, _)
return False
if count == 0:
log.debug('zero lines detected, blank input is not valid CSV')
return False
log.debug('%s CSV lines passed', count)
return True
def get_latest_failed_build(self):
log.info('getting latest failed build')
# gets 404 unless replacing the slash
url = 'https://api.travis-ci.org/repo/{repo}/builds'.format(repo=self.repo.replace('/', '%2F'))
# request returns blank without authorization header
req = self.request_handler.get(url, headers=self.headers)
if log.isEnabledFor(logging.DEBUG):
log.debug("\n%s", jsonpp(req.content))
try:
latest_build = self.parse_latest_failed_build(req.content)
except (KeyError, ValueError):
exception = traceback.format_exc().split('\n')[-2]
# this covers up the traceback info and makes it harder to debug
#raise UnknownError('failed to parse expected json response from Travis CI API: {0}'.format(exception))
qquit('UNKNOWN', 'failed to parse expected json response from Travis CI API: {0}. {1}'.
format(exception, support_msg_api()))
return latest_build
def req(self, method, url, *args, **kwargs):
if '://' not in url:
url = 'http://' + url
self.url = url
log.debug('%s %s', str(method).upper(), url)
req = None
if 'headers' in kwargs:
kwargs['headers']['User-Agent'] = get_topfile()
else:
kwargs['headers'] = {'User-Agent': get_topfile()}
try:
req = getattr(requests, method)(url, *args, **kwargs)
except requests.exceptions.RequestException as _:
self.exception_handler(_)
self.log_output(req)
self.process_req(req)
return req
def check_file(self, filename):
self.filename = filename
if self.filename == '-':
self.filename = '<STDIN>'
self.valid_ini_msg = '%s => INI OK' % self.filename
self.invalid_ini_msg = '%s => INI INVALID' % self.filename
if self.filename == '<STDIN>':
log.debug('ini stdin')
# TODO: should technically write to temp file to be able to seek(0) for print mode
self.check_ini(sys.stdin)
else:
log.debug('checking %s', self.filename)
try:
with open(self.filename) as iostream:
self.check_ini(iostream)
except IOError as _:
die("ERROR: %s" % _)
def run(self):
if not self.args:
self.usage('no Dockerfile / directory args given')
args = uniq_list_ordered(self.args)
self.branch_prefix = self.get_opt('branch_prefix')
if self.branch_prefix is not None:
validate_regex(self.branch_prefix, 'branch prefix')
self.branch_prefix = re.compile(self.branch_prefix)
for arg in args:
if not os.path.exists(arg):
print("'%s' not found" % arg)
sys.exit(ERRORS['WARNING'])
if os.path.isfile(arg):
log_option('file', arg)
elif os.path.isdir(arg):
log_option('directory', arg)
else:
die("path '%s' could not be determined as either a file or directory" % arg)
for arg in args:
self.check_git_branches_dockerfiles(arg)
log.info('Total Branches: %s', len(self.branches))
log.info('Selected Branches: %s', len(self.selected_branches))
log.info('Branches checked: %s', self.branches_checked)
log.info('Branches with Dockerfile checked: %s', len(self.branches_dockerfile_checked))
branches_skipped = len(self.branches_skipped)
if branches_skipped > 0:
log.warn('{0} branches skipped for not matching expected naming format'
.format(branches_skipped))
branches_not_checked = len(self.selected_branches) - len(self.branches_dockerfile_checked)
if branches_not_checked > 1:
log.warn('{0} branches not checked (no matching Dockerfile found?)'.format(branches_not_checked))
if log.isEnabledFor(logging.DEBUG):
log.debug('Branches with no corresponding Dockerfile found:\n%s',
'\n'.join(set(self.selected_branches) - set(self.branches_dockerfile_checked)))
log.info('{0} Dockerfiles checked'.format(len(self.dockerfiles_checked)))
branches_failed = len(self.branches_failed)
_ = '{0} Dockerfiles failed validation across {1} branches'.format(self.dockerfiles_failed, branches_failed)
if branches_failed > 0:
log.error(_)
else:
log.info(_)
if self.failed:
log.error('Dockerfile validation FAILED')
sys.exit(ERRORS['CRITICAL'])
log.info('Dockerfile validation SUCCEEDED')
def process_database(self, database, table_regex):
tables = []
table_count = 0
log.info("querying tables for database '%s'", database)
conn = self.connect(database)
with conn.cursor() as table_cursor:
try:
# doesn't support parameterized query quoting from dbapi spec
#table_cursor.execute('use %(database)s', {'database': database})
table_cursor.execute('use `{}`'.format(database))
table_cursor.execute('show tables')
except impala.error.HiveServer2Error as _:
log.error("error querying tables for database '%s': %s",
database, _)
if 'AuthorizationException' in str(_):
return
raise
for table_row in table_cursor:
table = table_row[0]
table_count += 1
if not table_regex.search(table):
log.debug("skipping database '%s' table '%s', does not match regex '%s'", \
database, table, self.table)
continue
tables.append(table)
log.info("%s/%s tables selected for database '%s'", len(tables),
table_count, database)
for table in tables:
try:
query = self.query.format(db='`{}`'.format(database),
table='`{}`'.format(table))
except KeyError as _:
if _ == 'db':
query = self.query.format(table='`{}`'.format(table))
else:
raise
try:
self.execute(conn, database, table, query)
self.table_count += 1
except Exception as _:
if self.ignore_errors:
log.error("database '%s' table '%s': %s", database, table,
_)
continue
raise
def list_keys(self, service_account_email):
log.debug("getting keys for service account '%s'",
service_account_email)
keys = self.service.projects()\
.serviceAccounts()\
.keys()\
.list(name='projects/-/serviceAccounts/' + service_account_email).execute()
for key in keys['keys']:
if key['keyType'] == 'SYSTEM_MANAGED':
continue
_id = key['name'].split('/')[-1]
created_date = key['validAfterTime']
expiry_date = key['validBeforeTime']
created_datetime = datetime.strptime(created_date,
"%Y-%m-%dT%H:%M:%SZ")
expiry_datetime = datetime.strptime(expiry_date,
"%Y-%m-%dT%H:%M:%SZ")
age_timedelta = datetime.utcnow() - created_datetime
age_days = int(age_timedelta.total_seconds() / 86400)
expired = False
if expiry_date == '9999-12-31T23:59:59Z':
expires_in_days = 'NEVER'
else:
expires_in_timedelta = expiry_datetime - datetime.utcnow()
expires_in_days = int(expires_in_timedelta.total_seconds() /
86400)
if expires_in_days < 1:
expired = True
if self.no_expiry and (expires_in_days != 'NEVER'):
continue
if self.expired and expired:
continue
if self.expires_within_days is not None and \
(expires_in_days == 'NEVER' or expires_in_days > self.expires_within_days):
continue
print(
'{id} {created} {expires} {age:4d} {expires_in:5s} {expired} {service_account}'
.format(id=_id,
created=created_date,
expires=expiry_date,
age=age_days,
expires_in=expires_in_days,
expired=expired,
service_account=service_account_email))
def check_dockerfile_arg(self, filename, branch):
log.debug('check_dockerfile_arg({0}, {1})'.format(filename, branch))
branch_base = str(branch).replace('-dev', '')
(branch_base, branch_versions) = self.branch_version(branch)
with open(filename) as filehandle:
version_index = 0
for line in filehandle:
#log.debug(line.strip())
# hack for Scala Java version, hacky but can't think of a better more generic way to do this right now
match = self.arg_regex.match(line.strip())
if match:
arg_var = match.group(1)
# this is too restrictive and prevents finding a lot of issues with
# more complex naming conventions for kafka, centos-java/scala etc
# instead we now expect ARG *_VERSION to be in the same order as the version numbers in branch name
#log.debug("checking arg group 1 '%s' == branch_base '%s'", argversion.group(1), branch_base)
#if self.normalize_name(arg_var) == self.normalize_name(branch_base).replace('-', '_'):
if version_index >= len(branch_versions):
return True
branch_version = branch_versions[version_index]
found_version = match.group(2)
if not self.check_version(filename=filename,
branch=branch,
branch_base=branch_base,
arg_var=arg_var,
found_version=found_version,
branch_version=branch_version):
return False
version_index += 1
elif branch_base == 'scala' \
and len(branch_versions) > 1 \
and branch_versions[1] is not None:
#log.debug('special scala condition checking for jdk version')
match = self.jdk_regex.match(line)
if match:
found_version = match.group(1)
#log.debug('found jdk version \'%s\'', found_version)
if not self.check_version(filename=filename,
branch=branch,
branch_base=branch_base,
arg_var=None,
found_version=found_version,
branch_version=branch_versions[version_index+1]):
return False
return True
def run(self):
iam = boto3.client('iam')
log.info('generating credentials report')
while True:
result = iam.generate_credential_report()
log.debug('%s', result)
if result['State'] == 'COMPLETE':
log.info('credentials report generated')
break
log.info('waiting for credentials report')
time.sleep(1)
try:
result = iam.get_credential_report()
except ClientError as _:
raise
csv_content = result['Content']
log.debug('%s', csv_content)
filehandle = StringIO(unicode(csv_content))
filehandle.seek(0)
csvreader = csv.reader(filehandle)
headers = next(csvreader)
assert headers[0] == 'user'
assert headers[4] == 'password_last_used'
assert headers[10] == 'access_key_1_last_used_date'
assert headers[15] == 'access_key_2_last_used_date'
self.now = datetime.utcnow()
found = False
for row in csvreader:
user = row[0]
if user != self.user:
continue
found = True
last_used_days = self.get_user_last_used_days(row)
if not found:
raise UnknownError('AWS user {} not found'.format(self.user))
if last_used_days <= self.days:
self.warning()
if last_used_days == 0:
self.msg = 'AWS user {} last used within the last day'.format(
self.user)
else:
self.msg = 'AWS user {} last used {} day{} ago'.format(
self.user, last_used_days, plural(last_used_days))
self.msg += ' | last_used_days={};0;;{}'.format(
last_used_days, self.days)
def process_message(self):
message = self.message
message = re.sub(
r'\s*(?:[\w\s]+?\s)?(?:OK|WARNING|CRITICAL|UNKNOWN)(?:\s[\w\s]+?)?\s*:\s*',
'', message, 1, re.I)
if re.search('^Hari Sekhon', message):
_ = re.search('^usage:', message, re.M)
if _:
log.debug(
'stripping off my extended plugin description header up to usage: options line'
+
'to make it more obvious that a usage error has occurred')
message = message[_.start():]
message = message.rstrip('\n')
message = re.sub(r'\r', '', message)
message = re.sub(r'\n', r' \\n ', message)
message = re.sub(r',\s*', '... ', message)
self.message = message
def get_version(self):
data = None
try:
conn = socket.create_connection((self.host, self.port), timeout=self.timeout/2)
conn.sendall('envi')
data = conn.recv(1024)
conn.close()
except socket.error as _:
qquit('CRITICAL', "Failed to connect to ZooKeeper at '{host}:{port}': "\
.format(host=self.host, port=self.port) + str(_))
version = None
log.debug(data.strip())
for line in data.split('\n'):
_ = self.version_line_regex.match(line)
if _:
version = _.group(1)
break
return version
def run(self):
log.info('testing AWS API call')
# there isn't really a .ping() type API endpoint so just connect to IAM and list users
iam = boto3.client('iam')
try:
_ = iam.list_users()
# just in case we get an iterator, consume it to flush out any error
collections.deque(_, maxlen=0)
if log.isEnabledFor(logging.DEBUG):
log.debug('\n\n%s', _)
log.debug('\n\n%s', jsonpp(_))
# pylint: disable=broad-except
except Exception as _:
if log.isEnabledFor(logging.DEBUG):
raise
else:
raise CriticalError(_)
self.msg = 'AWS API credentials OK'
def check_file(self, filename):
self.filename = filename
if self.filename == '-':
self.filename = '<STDIN>'
self.valid_csv_msg = '%s => CSV OK' % self.filename
self.invalid_csv_msg = '%s => CSV INVALID' % self.filename
if self.filename == '<STDIN>':
log.debug('checking stdin')
self.check_csv(sys.stdin)
else:
if self.is_excluded(filename):
return
log.debug('checking %s', self.filename)
try:
with open(self.filename) as iostream:
self.check_csv(iostream)
except IOError as _:
die("ERROR: %s" % _)
def process_table(self, table):
try:
table_handle = self.conn.table(table)
regions = table_handle.regions()
if len(regions) < 1:
qquit('UNKNOWN',
"no regions found for table '{0}'".format(table))
for region in regions:
log.debug("table '%s' region '%s'", table, region)
server = region['server_name']
self.server_region_counts[
server] = self.server_region_counts.get(server, 0)
self.server_region_counts[server] += 1
except (socket.timeout, ThriftException, HBaseIOError) as _:
qquit('CRITICAL', _)
except KeyError as _:
qquit('UNKNOWN',
'failed to process region information. ' + support_msg_api())
def parse(json_data):
try:
# it's already nicely layed out
#if log.isEnabledFor(logging.DEBUG):
# log.debug('%s', jsonpp(json_data))
compaction_queue_size = None
for bean in json_data['beans']:
if bean['name'] == 'Hadoop:service=HBase,name=RegionServer,sub=Server':
if log.isEnabledFor(logging.DEBUG):
log.debug('found RegionServer section:')
log.debug('%s', jsonpp(bean))
compaction_queue_size = bean['compactionQueueLength']
if not isInt(compaction_queue_size):
qquit('UNKNOWN', 'non-integer returned for compactionQueueLength! ' + support_msg_api())
return compaction_queue_size
except KeyError as _:
qquit('UNKNOWN', _ + ': failed to parse HBase Master jmx info. ' + support_msg_api())
qquit('UNKNOWN', 'RegionServer mbean not found, double check this is pointing to an HBase RegionServer')
def clean_detail(self):
detail = self.detail
detail = re.sub(
r'\s*(?:[\w\s]+?\s)?(?:OK|WARNING|CRITICAL|UNKNOWN)(?:\s[\w\s]+?)?\s*:\s*',
'', detail, 1, re.I)
if re.search('^Hari Sekhon', detail):
_ = re.search('^usage:', detail, re.M)
if _:
log.debug(
'stripping off my extended plugin description header up to usage: options line'
+
'to make it more obvious that a usage error has occurred')
detail = detail[_.start():]
detail = detail.rstrip('\n')
detail = re.sub(r'\r', '', detail)
detail = re.sub(r'\n', r' \\n ', detail)
detail = re.sub(r',\s*', '... ', detail)
self.detail = detail
def get_peers(content):
json_data = None
try:
json_data = json.loads(content)
except ValueError:
raise UnknownError("non-json data returned by consul: '%s'. %s" %
(content, support_msg_api()))
if not json_data:
raise CriticalError('no peers found, recently started?')
#if not json_data:
# raise UnknownError("blank list returned by consul! '%s'. %s" % (content, support_msg_api()))
if not isList(json_data):
raise UnknownError("non-list returned by consul: '%s'. %s" %
(content, support_msg_api()))
for peer in json_data:
log.debug('peer: {0}'.format(peer))
peers = uniq_list(json_data)
return peers
def check_table(self):
log.info('checking table \'%s\'', self.table)
if not self.conn.is_table_enabled(self.table):
qquit('CRITICAL', "table '{0}' is disabled!".format(self.table))
table_conn = self.conn.table(self.table)
families = table_conn.families()
self.num_column_families = len(families)
log.info('found %s column families: %s', self.num_column_families, families)
regions = table_conn.regions()
self.num_regions = len(regions)
log.info('found %s regions', self.num_regions)
if log.isEnabledFor(logging.DEBUG):
#log.debug('regions list:\n%s', '\n'.join([_['name'] for _ in regions]))
log.debug('regions list: \n%s', '\n'.join([str(_) for _ in regions]))
for column_family in sorted(families):
column = '{0}:{1}'.format(column_family, self.column_qualifier)
for region in regions:
self.check_region(table_conn, column, region)
def check_table(self):
log.info('checking table \'%s\'', self.table)
is_enabled = None
families = None
try:
is_enabled = self.conn.is_table_enabled(self.table)
log.info('enabled: %s', is_enabled)
table = self.conn.table(self.table)
families = table.families()
except HBaseIOError as _:
#if 'org.apache.hadoop.hbase.TableNotFoundException' in _.message:
if 'TableNotFoundException' in _.message:
qquit('CRITICAL', 'table \'{0}\' does not exist'.format(self.table))
else:
qquit('CRITICAL', _)
except (socket.error, socket.timeout, ThriftException) as _:
qquit('CRITICAL', _)
if log.isEnabledFor(logging.DEBUG):
log.debug('column families:\n%s', jsonpp(families))
if not families:
qquit('CRITICAL', 'failed to get column families for table \'{0}\''.format(self.table))
if not isDict(families):
qquit('UNKNOWN', 'column family info returned was not a dictionary! ' + support_msg_api())
num_families = len(families)
log.info('num families: %s', num_families)
self.msg = 'HBase table \'{0}\' is '.format(self.table)
if is_enabled:
self.msg += 'enabled, '
else:
self.critical()
self.msg += 'disabled! '
self.msg += '{0} column '.format(num_families)
if num_families == 1:
self.msg += 'family'
else:
self.msg += 'families'
self.check_thresholds(num_families)
self.msg += ' | num_column_families={0}'.format(num_families) + self.get_perf_thresholds(boundary='lower')
log.info('finished, closing connection')
self.conn.close()
def check_git_branches_dockerfiles(self, target):
gitroot = find_git_root(target)
if gitroot is None:
die('Failed to find git root for target {0}'.format(target))
log.debug("finding branches for target '{0}'".format(target))
repo = git.Repo(gitroot)
#branches = [str(x) for x in repo.refs if isinstance(x, git.refs.remote.RemoteReference)]
branches = [str(x) for x in repo.refs if isinstance(x, git.Head)]
branches = [x.split('/')[-1] for x in branches]
branches = set(branches)
branches = [x for x in branches if x not in ('HEAD', 'master')]
self.branches = branches
if self.branch_prefix is not None:
log.debug('restricting to branches matching branch prefix')
branches = [x for x in branches if self.branch_prefix.match(x)]
self.selected_branches = branches
#if log.isEnabledFor(logging.DEBUG):
log.debug('\n\nbranches for target %s:\n\n%s\n', target,
'\n'.join(branches))
# in Travis CI there is no original branch and master branch does not exist, so falling back to assuming master
# causes failure, better to not check out original branch if you don't know
#original_branch = 'master'
original_branch = None
try:
try:
original_branch = repo.active_branch.name
except TypeError as _:
pass
for branch in branches:
log.debug("checking branch '%s' Dockerfiles for target '%s'",
branch, target)
self.branches_checked += 1
try:
repo.git.checkout(branch)
except git.GitError as _:
die(_)
self.check_path(target, branch)
except Exception as _: # pylint: disable=broad-except
traceback.print_exc()
sys.exit(1)
finally:
if original_branch != None:
log.debug("checking out original branch '%s'", original_branch)
repo.git.checkout(original_branch)
def check(self, client):
log.info('getting Docker version')
_ = client.version()
if log.isEnabledFor(logging.DEBUG):
log.debug((jsonpp(_)))
version = _['Version']
if not isVersionLax(version):
raise UnknownError('Docker version unrecognized \'{}\'. {}'\
.format(version, support_msg_api()))
self.msg = 'Docker version = {}'.format(version)
if self.expected is not None:
log.info("verifying version against expected regex '%s'", self.expected)
if re.match(self.expected, str(version)):
log.info('version regex matches retrieved version')
else:
log.info('version regex does not match retrieved version')
self.msg += " (expected '{}')".format(self.expected)
self.critical()
self.msg += ', API version = {}'.format(_['ApiVersion'])
def run_host(self, host, url):
log.info('querying %s', host)
req = RequestHandler().get(url)
json_data = json.loads(req.text)
uptime = None
beans = json_data['beans']
if self.since_uptime:
for bean in beans:
if bean['name'] == 'java.lang:type=Runtime':
uptime = int(bean['Uptime'] / 1000)
break
if not uptime:
raise UnknownError("failed to find uptime in JMX stats for host '{}'. {}"\
.format(host, support_msg_api()))
for bean in beans:
log.debug('processing Regions bean')
if bean['name'] == 'Hadoop:service=HBase,name=RegionServer,sub=Regions':
self.process_bean(host, bean, uptime)
self.print_stats(host)
def fetch(self, url_suffix):
err = ''
try:
response = self.get(url_suffix)
except requests.exceptions.RequestException as _:
err = "failed to fetch Ambari Blueprint from '%s': %s" % (self.url,
_)
# log.critical(err)
qquit('CRITICAL', err)
json_data = json.loads(response)
if log.isEnabledFor(logging.DEBUG):
log.debug("blueprint = " + jsonpp(json_data))
try:
del json_data['href']
log.debug(
"stripped href as it's not valid if re-submitting the blueprint to Ambari"
)
except KeyError as _:
pass
# Ambari 2.1.3 supports this according to:
# https://cwiki.apache.org/confluence/display/AMBARI/Blueprints#Blueprints-ClusterCreationTemplateStructure
# json_data['config_recommendation_strategy'] = 'NEVER_APPLY' # default
# json_data['config_recommendation_strategy'] = 'ONLY_STACK_DEFAULTS_APPLY'
# json_data['config_recommendation_strategy'] = 'ALWAYS_APPLY'
if self.strip_config:
log.info(
'stripping out config sections of blueprints to make more generic'
)
try:
del json_data['configurations']
for hostgroup in json_data['host_groups']:
del hostgroup['configurations']
except KeyError as _:
pass
try:
json_data['host_groups'] = list_sort_dicts_by_value(
json_data['host_groups'], 'name')
for hostgroup in json_data['host_groups']:
hostgroup['components'] = list_sort_dicts_by_value(
hostgroup['components'], 'name')
except KeyError as _:
qquit('CRITICAL', 'failed to sort blueprint: %s' % _)
return jsonpp(json_data)
def run(self): # pylint: disable=no-self-use
linux_only()
regex = re.compile(r'^HugePages_Total:\s+(\d+)\s*$')
hugepages_total = None
with open('/proc/meminfo') as meminfo:
for line in meminfo:
if 'HugePage' in line:
log.debug(line)
match = regex.match(line)
if match:
hugepages_total = int(match.group(1)) # protected by regex
break
if hugepages_total is None:
raise UnknownError('HugePages Total not found in /proc/meminfo. {}'.format(support_msg()))
if hugepages_total == 0:
self.msg += 'disabled'
else:
raise CriticalError(' Huge Pages = enabled. This should be disabled for Big Data ' +
'systems such as Hadoop / MongoDB for performance reasons etc...')
