Author: phunt
Date: Wed Jul 14 06:39:32 2010
New Revision: 963957
URL: http://svn.apache.org/viewvc?rev=963957&view=rev
Log:
ZOOKEEPER-799. Add tools and recipes for monitoring as a contrib
Added:
hadoop/zookeeper/trunk/src/contrib/monitoring/
hadoop/zookeeper/trunk/src/contrib/monitoring/JMX-RESOURCES
hadoop/zookeeper/trunk/src/contrib/monitoring/README
hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/
hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/README
hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py (with
props)
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/README
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/Screenshot.png
(with props)
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/modpython.conf
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper.pyconf
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper_ganglia.py
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/README.txt
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot-1.png
(with props)
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot.png (with
props)
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/hostgroups.cfg
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/services.cfg
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/zookeeper.cfg
hadoop/zookeeper/trunk/src/contrib/monitoring/test.py (with props)
Modified:
hadoop/zookeeper/trunk/CHANGES.txt
Modified: hadoop/zookeeper/trunk/CHANGES.txt
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/CHANGES.txt?rev=963957&r1=963956&r2=963957&view=diff
==============================================================================
--- hadoop/zookeeper/trunk/CHANGES.txt (original)
+++ hadoop/zookeeper/trunk/CHANGES.txt Wed Jul 14 06:39:32 2010
@@ -33,7 +33,7 @@ BUGFIXES:
do not handle windows path correctly (Vishal K via phunt)
ZOOKEEPER-754. numerous misspellings "succesfully"
- (Savu Andrei via phunt)
+ (Andrei Savu via phunt)
ZOOKEEPER-749. OSGi metadata not included in binary only jar (phunt
via henryr)
@@ -97,10 +97,14 @@ NEW FEATURES:
ZOOKEEPER-773. Log visualisation (Ivan Kelly via phunt)
- ZOOKEEPER-744. Add monitoring four-letter word (Savu Andrei via phunt)
+ ZOOKEEPER-744. Add monitoring four-letter word (Andrei Savu via phunt)
ZOOKEEPER-712. Bookie recovery. (erwin tam via breed)
+ ZOOKEEPER-799. Add tools and recipes for monitoring as a contrib
+ (Andrei Savu via phunt)
+
+
Release 3.3.0 - 2010-03-24
Non-backward compatible changes:
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/JMX-RESOURCES
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/JMX-RESOURCES?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/JMX-RESOURCES (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/JMX-RESOURCES Wed Jul 14
06:39:32 2010
@@ -0,0 +1,38 @@
+
+Resources for monitoring ZooKeeper using JMX
+--------------------------------------------
+
+JMX/REST Bridge
+---------------
+
+http://code.google.com/p/polarrose-jmx-rest-bridge/
+
+"Simple Java Web Application that exposes JMX servers through HTTP. This was
written so that external tools can easily query JMX attributes of Java
applications. More specifically, this was written to allow Cacti to generate
fancy graphs of ActiveMQ instances."
+
+JMXetric
+--------
+
+http://code.google.com/p/jmxetric/
+
+"JMXetric is a 100% java, configurable JVM agent that periodically polls MBean
attributes and reports their values to Ganglia."
+
+jmxquery
+--------
+
+http://code.google.com/p/jmxquery/
+
+"a plugin for nagios to check jmx"
+
+check_jmx
+---------
+
+http://exchange.nagios.org/directory/Plugins/Java-Applications-and-Servers/check_jmx/details
+
+
+jmx2snmp
+--------
+
+http://github.com/tcurdt/jmx2snmp
+
+Expose application JMX properties via SNMP
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/README
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/README?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/README (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/README Wed Jul 14 06:39:32
2010
@@ -0,0 +1,79 @@
+
+Tools and Recipes for ZooKeeper Monitoring
+------------------------------------------
+
+How To Monitor
+--------------
+
+A ZooKeeper cluster can be monitored in two ways:
+ 1. by using the 'mntr' 4letterword command
+ 2. by using JMX to query the MBeans
+
+This repo contains tools and recipes for monitoring ZooKeeper using the first
method.
+
+Check the file JMX-RESOURCE for some links to resources that could help you
monitor a ZooKeeper cluster using the JMX interface.
+
+Requirements
+------------
+
+ZooKeeper 3.4.0 or later or you can apply ZOOKEEPER-744 patch over the latest
3.3.x release.
+The server should understand the 'mntr' 4letterword command.
+
+$ echo 'mntr' | nc localhost 2181
+zk_version 3.4.0--1, built on 06/19/2010 15:07 GMT
+zk_avg_latency 141
+zk_max_latency 1788
+zk_min_latency 0
+zk_packets_received 385466
+zk_packets_sent 435364
+zk_outstanding_requests 0
+zk_server_state follower
+zk_znode_count 5
+zk_watch_count 0
+zk_ephemerals_count 0
+zk_approximate_data_size 41
+zk_open_file_descriptor_count 20
+zk_max_file_descriptor_count 1024
+
+Python 2.6 (maybe it works on previous version but it's not tested yet).
+
+In a nutshell
+-------------
+
+All you need is check_zookeeper.py It has no external dependencies.
+
+
+*** On Nagios call the script like this:
+
+./check_zookeeper.py -o nagios -s "<server-or-list-of-servers>" -k <key> -w
<warning> -c <critical>
+
+
+*** On Cacti define a custom data input method using the script like this:
+
+./check_zookeeper.py -o cacti -s "<list-of-servers>" -k <key> --leader
+
+-- outputs a single value for the given key fetched from the cluster leader
+
+OR
+
+./check_zookeeper.py -o cacti -s "<list-of-servers>" -k <key>
+
+-- outputs multiple values on for each cluster node
+ex: localhost_2182:0 localhost_2183:0 localhost_2181:0 localhost_2184:0
localhost_2185:0
+
+*** On Ganglia:
+
+install the plugin found in the ganglia/ subfolder OR
+
+./check_zookeeper.py -o ganglia -s "<current-zookeeper-node>"
+
+it will use gmetric to send zookeeper node status data.
+
+
+Check the subfolders for configuration details and samples for each platform.
+
+ZooKeeper 4letterwords Commands
+-------------------------------
+
+http://hadoop.apache.org/zookeeper/docs/current/zookeeperAdmin.html#sc_zkCommands
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/README
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/README?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/README (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/cacti/README Wed Jul 14
06:39:32 2010
@@ -0,0 +1,41 @@
+
+Recipes for ZooKeeper monitoring using Cacti
+--------------------------------------------
+
+Cacti install guide: https://help.ubuntu.com/community/Cacti
+
+Cacti Manual: http://www.cacti.net/downloads/docs/html/
+PDF version: http://www.cacti.net/downloads/docs/pdf/manual.pdf
+
+Check Chapter 16: Simplest Method of Going from Script to Graph
+ http://www.cacti.net/downloads/docs/html/how_to.html#SCRIPT_TO_GRAPH
+
+WARNING: I have wrote these instructions while installing and configuring the
plugin on my desktop computer running Ubuntu 9.10. I've installed Cacti using
apt-get.
+
+WARNING: I'm going to make the assumption that you know how to work with Cacti
and how to setup Data Input Methods for custom scripts. I'm also going to
assume that you have already installed Cacti and everything works as expected.
+
+You can extend the Cacti's data gathering functionality through external
scripts. Cacti comes with a number of scripts out of the box wich are localted
in the scripts/ directory.
+
+
+The check_zookeeper.py script can be used a custom data input method for
Cacti.
+
+Single value (check cluster status by sending queries to the leader):
+---------------------------------------------------------------------
+
+python <path_cacti>scripts/check_zookeeper.py -s
"localhost:2181,localhost:2182,localhost:2183,localhost:2184,localhost:2185" -k
<key> -o cacti --leader
+
+When you will call the script this way it will about a single value
representing the value attached to this <key>.
+
+
+Multiple values (one for each cluster node):
+--------------------------------------------
+
+python <path_cacti>scripts/check_zookeeper.py -s
"localhost:2181,localhost:2182,localhost:2183,localhost:2184,localhost:2185" -k
<key> -o cacti
+
+Output:
+localhost_2182:0 localhost_2183:0 localhost_2181:0 localhost_2184:0
localhost_2185:0
+
+
+TBD: Step by step guide
+
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py Wed Jul 14
06:39:32 2010
@@ -0,0 +1,338 @@
+#! /usr/bin/env python
+""" Check Zookeeper Cluster
+
+Generic monitoring script that could be used with multiple platforms (Ganglia,
Nagios, Cacti).
+
+It requires ZooKeeper 3.4.0 or greater. The script needs the 'mntr' 4letter
word
+command (patch ZOOKEEPER-744) that was now commited to the trunk.
+The script also works with ZooKeeper 3.3.x but in a limited way.
+"""
+
+import sys
+import socket
+import logging
+import re
+import subprocess
+
+from StringIO import StringIO
+from optparse import OptionParser, OptionGroup
+
+__version__ = (0, 1, 0)
+
+log = logging.getLogger()
+logging.basicConfig(level=logging.ERROR)
+
+class NagiosHandler(object):
+
+ @classmethod
+ def register_options(cls, parser):
+ group = OptionGroup(parser, 'Nagios specific options')
+
+ group.add_option('-w', '--warning', dest='warning')
+ group.add_option('-c', '--critical', dest='critical')
+
+ parser.add_option_group(group)
+
+ def analyze(self, opts, cluster_stats):
+ try:
+ warning = int(opts.warning)
+ critical = int(opts.critical)
+
+ except (TypeError, ValueError):
+ print >>sys.stderr, 'Invalid values for "warning" and "critical".'
+ return 2
+
+ if opts.key is None:
+ print >>sys.stderr, 'You should specify a key name.'
+ return 2
+
+ warning_state, critical_state, values = [], [], []
+ for host, stats in cluster_stats.items():
+ if opts.key in stats:
+
+ value = stats[opts.key]
+ values.append('%s=%s;%s;%s' % (host, value, warning, critical))
+
+ if warning >= value > critical or warning <= value < critical:
+ warning_state.append(host)
+
+ elif (warning < critical and critical <= value) or (warning >
critical and critical >= value):
+ critical_state.append(host)
+
+ values = ' '.join(values)
+ if critical_state:
+ print 'Critical "%s" %s!|%s' % (opts.key, ',
'.join(critical_state), values)
+ return 2
+
+ elif warning_state:
+ print 'Warning "%s" %s!|%s' % (opts.key, ', '.join(warning_state),
values)
+ return 1
+
+ else:
+ print 'Ok "%s"!|%s' % (opts.key, values)
+ return 0
+
+class CactiHandler(object):
+
+ @classmethod
+ def register_options(cls, parser):
+ group = OptionGroup(parser, 'Cacti specific options')
+
+ group.add_option('-l', '--leader', dest='leader',
+ action="store_true", help="only query the cluster leader")
+
+ parser.add_option_group(group)
+
+ def analyze(self, opts, cluster_stats):
+ if opts.key is None:
+ print >>sys.stderr, 'The key name is mandatory.'
+ return 1
+
+ if opts.leader is True:
+ try:
+ leader = [x for x in cluster_stats.values() \
+ if x.get('zk_server_state', '') == 'leader'][0]
+
+ except IndexError:
+ print >>sys.stderr, 'No leader found.'
+ return 3
+
+ if opts.key in leader:
+ print leader[opts.key]
+ return 0
+
+ else:
+ print >>sys.stderr, 'Unknown key: "%s"' % opts.key
+ return 2
+ else:
+ for host, stats in cluster_stats.items():
+ if opts.key not in stats:
+ continue
+
+ host = host.replace(':', '_')
+ print '%s:%s' % (host, stats[opts.key]),
+
+
+class GangliaHandler(object):
+
+ @classmethod
+ def register_options(cls, parser):
+ group = OptionGroup(parser, 'Ganglia specific options')
+
+ group.add_option('-g', '--gmetric', dest='gmetric',
+ default='/usr/bin/gmetric', help='ganglia gmetric binary '\
+ 'location: /usr/bin/gmetric')
+
+ parser.add_option_group(group)
+
+ def call(self, *args, **kwargs):
+ subprocess.call(*args, **kwargs)
+
+ def analyze(self, opts, cluster_stats):
+ if len(cluster_stats) != 1:
+ print >>sys.stderr, 'Only allowed to monitor a single node.'
+ return 1
+
+ for host, stats in cluster_stats.items():
+ for k, v in stats.items():
+ try:
+ self.call([opts.gmetric, '-n', k, '-v', str(int(v)), '-t',
'uint32'])
+ except (TypeError, ValueError):
+ pass
+
+class ZooKeeperServer(object):
+
+ def __init__(self, host='localhost', port='2181', timeout=1):
+ self._address = (host, int(port))
+ self._timeout = timeout
+
+ def get_stats(self):
+ """ Get ZooKeeper server stats as a map """
+ data = self._send_cmd('mntr')
+ if data:
+ return self._parse(data)
+ else:
+ data = self._send_cmd('stat')
+ return self._parse_stat(data)
+
+ def _create_socket(self):
+ return socket.socket()
+
+ def _send_cmd(self, cmd):
+ """ Send a 4letter word command to the server """
+ s = self._create_socket()
+ s.settimeout(self._timeout)
+
+ s.connect(self._address)
+ s.send(cmd)
+
+ data = s.recv(2048)
+ s.close()
+
+ return data
+
+ def _parse(self, data):
+ """ Parse the output from the 'mntr' 4letter word command """
+ h = StringIO(data)
+
+ result = {}
+ for line in h.readlines():
+ try:
+ key, value = self._parse_line(line)
+ result[key] = value
+ except ValueError:
+ pass # ignore broken lines
+
+ return result
+
+ def _parse_stat(self, data):
+ """ Parse the output from the 'stat' 4letter word command """
+ h = StringIO(data)
+
+ result = {}
+
+ version = h.readline()
+ if version:
+ result['zk_version'] = version[version.index(':')+1:].strip()
+
+ # skip all lines until we find the empty one
+ while h.readline().strip(): pass
+
+ for line in h.readlines():
+ m = re.match('Latency min/avg/max: (\d+)/(\d+)/(\d+)', line)
+ if m is not None:
+ result['zk_min_latency'] = int(m.group(1))
+ result['zk_avg_latency'] = int(m.group(2))
+ result['zk_max_latency'] = int(m.group(3))
+ continue
+
+ m = re.match('Received: (\d+)', line)
+ if m is not None:
+ result['zk_packets_received'] = int(m.group(1))
+ continue
+
+ m = re.match('Sent: (\d+)', line)
+ if m is not None:
+ result['zk_packets_sent'] = int(m.group(1))
+ continue
+
+ m = re.match('Outstanding: (\d+)', line)
+ if m is not None:
+ result['zk_outstanding_requests'] = int(m.group(1))
+ continue
+
+ m = re.match('Mode: (.*)', line)
+ if m is not None:
+ result['zk_server_state'] = m.group(1)
+ continue
+
+ m = re.match('Node count: (\d+)', line)
+ if m is not None:
+ result['zk_znode_count'] = int(m.group(1))
+ continue
+
+ return result
+
+ def _parse_line(self, line):
+ try:
+ key, value = map(str.strip, line.split('\t'))
+ except ValueError:
+ raise ValueError('Found invalid line: %s' % line)
+
+ if not key:
+ raise ValueError('The key is mandatory and should not be empty')
+
+ try:
+ value = int(value)
+ except (TypeError, ValueError):
+ pass
+
+ return key, value
+
+def main():
+ opts, args = parse_cli()
+
+ cluster_stats = get_cluster_stats(opts.servers)
+ if opts.output is None:
+ dump_stats(cluster_stats)
+ return 0
+
+ handler = create_handler(opts.output)
+ if handler is None:
+ log.error('undefined handler: %s' % opts.output)
+ sys.exit(1)
+
+ return handler.analyze(opts, cluster_stats)
+
+def create_handler(name):
+ """ Return an instance of a platform specific analyzer """
+ try:
+ return globals()['%sHandler' % name.capitalize()]()
+ except KeyError:
+ return None
+
+def get_all_handlers():
+ """ Get a list containing all the platform specific analyzers """
+ return [NagiosHandler, CactiHandler, GangliaHandler]
+
+def dump_stats(cluster_stats):
+ """ Dump cluster statistics in an user friendly format """
+ for server, stats in cluster_stats.items():
+ print 'Server:', server
+
+ for key, value in stats.items():
+ print "%30s" % key, ' ', value
+ print
+
+def get_cluster_stats(servers):
+ """ Get stats for all the servers in the cluster """
+ stats = {}
+ for host, port in servers:
+ try:
+ zk = ZooKeeperServer(host, port)
+ stats["%s:%s" % (host, port)] = zk.get_stats()
+
+ except socket.error, e:
+ # ignore because the cluster can still work even
+ # if some servers fail completely
+
+ # this error should be also visible in a variable
+ # exposed by the server in the statistics
+
+ logging.info('unable to connect to server '\
+ '"%s" on port "%s"' % (host, port))
+
+ return stats
+
+
+def get_version():
+ return '.'.join(map(str, __version__))
+
+
+def parse_cli():
+ parser = OptionParser(usage='./check_zookeeper.py <options>',
version=get_version())
+
+ parser.add_option('-s', '--servers', dest='servers',
+ help='a list of SERVERS', metavar='SERVERS')
+
+ parser.add_option('-o', '--output', dest='output',
+ help='output HANDLER: nagios, ganglia, cacti', metavar='HANDLER')
+
+ parser.add_option('-k', '--key', dest='key')
+
+ for handler in get_all_handlers():
+ handler.register_options(parser)
+
+ opts, args = parser.parse_args()
+
+ if opts.servers is None:
+ parser.error('The list of servers is mandatory')
+
+ opts.servers = [s.split(':') for s in opts.servers.split(',')]
+
+ return (opts, args)
+
+
+if __name__ == '__main__':
+ sys.exit(main())
+
Propchange: hadoop/zookeeper/trunk/src/contrib/monitoring/check_zookeeper.py
------------------------------------------------------------------------------
svn:executable = *
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/README
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/README?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/README (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/README Wed Jul 14
06:39:32 2010
@@ -0,0 +1,33 @@
+
+Recipes for ZooKeeper monitoring using Ganglia
+----------------------------------------------
+
+Ganglia Install guide:
http://sourceforge.net/apps/trac/ganglia/wiki/Ganglia%203.1.x%20Installation%20and%20Configuration
+
+Gmond configuration:
http://sourceforge.net/apps/trac/ganglia/wiki/Gmond%203.1.x%20General%20Configuration
+
+WARNING: I have wrote these instructions while installing and configuring the
plugin on my desktop computer running Ubuntu 9.10. I've installed Ganglia using
apt-get.
+
+WARNING: I'm going to make the assumption that you know how to work with
Ganglia. I'm also going to assume that you have already installed Gangli and
everything works as expected.
+
+You can monitoring ZooKeeper using Ganglia in two ways:
+
+1. Using a python module:
+
+ WARNING! The python module only works with Ganglia 3.1.x
+
+ a. enable python modules: you can find instructions in modpython.confg
+ b. copy zookeeper.pyconf in /etc/ganglia/conf.d/
+ c. copy zookeeper_ganglia.py in /usr/lib/ganglia/python_plugins
+ d. restart the ganglia-monitor
+
+ This is the recommended way!
+
+2. OR Using check_zookeeper.py and gmetric:
+
+ Monitoring ZooKeeper using Ganglia is a simple as calling:
+
+ ./check_zookeeper.py -o ganglia -s localhost:2181
+
+ on each of the ZooKeeper cluster nodes. I'm making the assumption that you
have already configured gmond and installed gmetric on each node.
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/Screenshot.png
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/Screenshot.png?rev=963957&view=auto
==============================================================================
Binary file - no diff available.
Propchange: hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/Screenshot.png
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/modpython.conf
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/modpython.conf?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/modpython.conf (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/modpython.conf Wed
Jul 14 06:39:32 2010
@@ -0,0 +1,13 @@
+
+/* Update gmond.conf */
+
+modules {
+ module {
+ name = "python_module"
+ path = "/usr/lib/ganglia/modpython.so"
+ params = "/usr/lib/ganglia/python_modules"
+ }
+}
+
+include ('/etc/ganglia/conf.d/*.pyconf')
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper.pyconf
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper.pyconf?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper.pyconf
(added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper.pyconf Wed
Jul 14 06:39:32 2010
@@ -0,0 +1,34 @@
+
+/* Update /etc/ganglia/gmond.conf with the content of this file. */
+
+/* ATTENTION: Change the host and the port to meet your setup. */
+
+modules {
+ module {
+ name = "zookeeper_ganglia"
+ language = "python"
+ param host { value = "127.0.0.1" }
+ param port { value = 2181 }
+ }
+}
+
+collection_group {
+ collect_every = 20
+ time_threshold = 60
+ metric { name = "zk_avg_latency" }
+ metric { name = "zk_max_latency" }
+ metric { name = "zk_min_latency" }
+ metric { name = "zk_packets_received" }
+ metric { name = "zk_packets_sent" }
+ metric { name = "zk_outstanding_requests" }
+ metric { name = "zk_znode_count" }
+ metric { name = "zk_watch_count" }
+ metric { name = "zk_ephemerals_count" }
+ metric { name = "zk_approximate_data_size" }
+ metric { name = "zk_open_file_descriptor_count" }
+ metric { name = "zk_max_file_descriptor_count" }
+ metric { name = "zk_followers" }
+ metric { name = "zk_synced_followers" }
+ metric { name = "zk_pending_syncs" }
+}
+
Added:
hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper_ganglia.py
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper_ganglia.py?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper_ganglia.py
(added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/ganglia/zookeeper_ganglia.py
Wed Jul 14 06:39:32 2010
@@ -0,0 +1,194 @@
+""" Python Ganglia Module for ZooKeeper monitoring
+
+Inspired by: http://gist.github.com/448007
+
+Copy this file to /usr/lib/ganglia/python_plugins
+
+"""
+
+import sys
+import socket
+import time
+import re
+
+from StringIO import StringIO
+
+TIME_BETWEEN_QUERIES = 20
+
+class ZooKeeperServer(object):
+
+ def __init__(self, host='localhost', port='2181', timeout=1):
+ self._address = (host, int(port))
+ self._timeout = timeout
+
+ def get_stats(self):
+ """ Get ZooKeeper server stats as a map """
+ data = self._send_cmd('mntr')
+ if data:
+ return self._parse(data)
+ else:
+ data = self._send_cmd('stat')
+ return self._parse_stat(data)
+
+ def _create_socket(self):
+ return socket.socket()
+
+ def _send_cmd(self, cmd):
+ """ Send a 4letter word command to the server """
+ s = self._create_socket()
+ s.settimeout(self._timeout)
+
+ s.connect(self._address)
+ s.send(cmd)
+
+ data = s.recv(2048)
+ s.close()
+
+ return data
+
+ def _parse(self, data):
+ """ Parse the output from the 'mntr' 4letter word command """
+ h = StringIO(data)
+
+ result = {}
+ for line in h.readlines():
+ try:
+ key, value = self._parse_line(line)
+ result[key] = value
+ except ValueError:
+ pass # ignore broken lines
+
+ return result
+
+ def _parse_stat(self, data):
+ """ Parse the output from the 'stat' 4letter word command """
+ h = StringIO(data)
+
+ result = {}
+
+ version = h.readline()
+ if version:
+ result['zk_version'] = version[version.index(':')+1:].strip()
+
+ # skip all lines until we find the empty one
+ while h.readline().strip(): pass
+
+ for line in h.readlines():
+ m = re.match('Latency min/avg/max: (\d+)/(\d+)/(\d+)', line)
+ if m is not None:
+ result['zk_min_latency'] = int(m.group(1))
+ result['zk_avg_latency'] = int(m.group(2))
+ result['zk_max_latency'] = int(m.group(3))
+ continue
+
+ m = re.match('Received: (\d+)', line)
+ if m is not None:
+ result['zk_packets_received'] = int(m.group(1))
+ continue
+
+ m = re.match('Sent: (\d+)', line)
+ if m is not None:
+ result['zk_packets_sent'] = int(m.group(1))
+ continue
+
+ m = re.match('Outstanding: (\d+)', line)
+ if m is not None:
+ result['zk_outstanding_requests'] = int(m.group(1))
+ continue
+
+ m = re.match('Mode: (.*)', line)
+ if m is not None:
+ result['zk_server_state'] = m.group(1)
+ continue
+
+ m = re.match('Node count: (\d+)', line)
+ if m is not None:
+ result['zk_znode_count'] = int(m.group(1))
+ continue
+
+ return result
+
+ def _parse_line(self, line):
+ try:
+ key, value = map(str.strip, line.split('\t'))
+ except ValueError:
+ raise ValueError('Found invalid line: %s' % line)
+
+ if not key:
+ raise ValueError('The key is mandatory and should not be empty')
+
+ try:
+ value = int(value)
+ except (TypeError, ValueError):
+ pass
+
+ return key, value
+
+def metric_handler(name):
+ if time.time() - metric_handler.timestamp > TIME_BETWEEN_QUERIES:
+ zk = ZooKeeperServer(metric_handler.host, metric_handler.port, 5)
+ try:
+ metric_handler.info = zk.get_stats()
+ except Exception, e:
+ print >>sys.stderr, e
+ metric_handler.info = {}
+
+ return metric_handler.info.get(name, 0)
+
+def metric_init(params=None):
+ params = params or {}
+
+ metric_handler.host = params.get('host', 'localhost')
+ metric_handler.port = int(params.get('port', 2181))
+ metric_handler.timestamp = 0
+
+ metrics = {
+ 'zk_avg_latency': {'units': 'ms'},
+ 'zk_max_latency': {'units': 'ms'},
+ 'zk_min_latency': {'units': 'ms'},
+ 'zk_packets_received': {
+ 'units': 'packets',
+ 'slope': 'positive'
+ },
+ 'zk_packets_sent': {
+ 'units': 'packets',
+ 'slope': 'positive'
+ },
+ 'zk_outstanding_requests': {'units': 'connections'},
+ 'zk_znode_count': {'units': 'znodes'},
+ 'zk_watch_count': {'units': 'watches'},
+ 'zk_ephemerals_count': {'units': 'znodes'},
+ 'zk_approximate_data_size': {'units': 'bytes'},
+ 'zk_open_file_descriptor_count': {'units': 'descriptors'},
+ 'zk_max_file_descriptor_count': {'units': 'descriptors'},
+ 'zk_followers': {'units': 'nodes'},
+ 'zk_synced_followers': {'units': 'nodes'},
+ 'zk_pending_syncs': {'units': 'syncs'}
+ }
+ metric_handler.descriptors = {}
+ for name, updates in metrics.iteritems():
+ descriptor = {
+ 'name': name,
+ 'call_back': metric_handler,
+ 'time_max': 90,
+ 'value_type': 'int',
+ 'units': '',
+ 'slope': 'both',
+ 'format': '%d',
+ 'groups': 'zookeeper',
+ }
+ descriptor.update(updates)
+ metric_handler.descriptors[name] = descriptor
+
+ return metric_handler.descriptors.values()
+
+def metric_cleanup():
+ pass
+
+
+if __name__ == '__main__':
+ ds = metric_init({'host':'localhost', 'port': '2181'})
+ for d in ds:
+ print "%s=%s" % (d['name'], metric_handler(d['name']))
+
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/README.txt
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/README.txt?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/README.txt (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/README.txt Wed Jul 14
06:39:32 2010
@@ -0,0 +1,71 @@
+
+Configuration Recipe for monitoring ZooKeeper using Nagios
+----------------------------------------------------------
+
+I will start by making the assumption that you already have an working Nagios
install.
+
+WARNING: I have wrote these instructions while installing and configuring the
plugin on my desktop computer running Ubuntu 9.10. I've installed Nagios using
apt-get.
+
+WARNING: You should customize the config files as suggested in order to match
your Nagios and Zookeeper install.
+
+WARNING: This README assumes you know how to configure Nagios and how it
works.
+
+WARNING: You should customize the warning and critical levels on service
checks to meet your own needs.
+
+1. Install the plugin
+
+$ cp check_zookeeper.py /usr/lib/nagios/plugins/
+
+2. Install the new commands
+
+$ cp zookeeper.cfg /etc/nagios-plugins/config
+
+3. Update the list of servers in zookeeper.cfg for the command
'check_zookeeper' and update the port for the command 'check_zk_node' (default:
2181)
+
+4. Create a virtual host in Nagios used for monitoring the cluster as a whole
-OR- Create a hostgroup named 'zookeeper-servers' and add all the zookeeper
cluster nodes.
+
+5. Define service checks like I have ilustrated bellow or just use the
provided definitions.
+
+define service {
+ use generic-service
+ host_name zookeeper-cluster
+ service_description ...
+ check_command
check_zookeeper!<exported-var>!<warning-level>!<critical-level>
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Open_File_Descriptors_Count
+ check_command check_zk_node!<exported-var>!<warning-level>!<critical-level>
+}
+
+Ex:
+
+a. check the number of open file descriptors
+
+define service{
+ use generic-service
+ host_name zookeeper-cluster
+ service_description ZK_Open_File_Descriptor_Count
+ check_command check_zookeeper!zk_open_file_descriptor_count!500!800
+}
+
+b. check the number of ephemerals nodes
+
+define service {
+ use generic-service
+ host_name localhost
+ service_description ZK_Ephemerals_Count
+ check_command check_zookeeper!zk_ephemerals_count!10000!100000
+}
+
+c. check the number of open file descriptors for each host in the group
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Open_File_Descriptors_Count
+ check_command check_zk_node!zk_open_file_descriptor_count!500!800
+}
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot-1.png
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot-1.png?rev=963957&view=auto
==============================================================================
Binary file - no diff available.
Propchange:
hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot-1.png
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot.png
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot.png?rev=963957&view=auto
==============================================================================
Binary file - no diff available.
Propchange: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/Screenshot.png
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/hostgroups.cfg
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/hostgroups.cfg?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/hostgroups.cfg (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/hostgroups.cfg Wed Jul
14 06:39:32 2010
@@ -0,0 +1,10 @@
+
+# A group containing all the ZooKeeper nodes
+
+define hostgroup {
+ hostgroup_name zookeeper-servers
+ alias ZooKeeper Servers
+ members localhost
+}
+
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/services.cfg
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/services.cfg?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/services.cfg (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/services.cfg Wed Jul
14 06:39:32 2010
@@ -0,0 +1,51 @@
+# ZooKeeper Node specific services
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Open_File_Descriptors_Count
+ check_command check_zk_node!zk_open_file_descriptor_count!500!800
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Ephemerals_Count
+ check_command check_zk_node!zk_ephemerals_count!10000!100000
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Avg_Latency
+ check_command check_zk_node!zk_avg_latency!500!1000
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Max_Latency
+ check_command check_zk_node!zk_max_latency!1000!2000
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Min_Latency
+ check_command check_zk_node!zk_min_latency!500!1000
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Outstanding_Requests
+ check_command check_zk_node!zk_outstanding_requests!20!50
+}
+
+define service {
+ hostgroup_name zookeeper-servers
+ use generic-service
+ service_description ZK_Watch_Count
+ check_command check_zk_node!zk_watch_count!100!500
+}
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/zookeeper.cfg
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/zookeeper.cfg?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/zookeeper.cfg (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/nagios/zookeeper.cfg Wed Jul
14 06:39:32 2010
@@ -0,0 +1,15 @@
+
+# 'check_zookeeper' command definition
+define command {
+ command_name check_zookeeper
+ command_line /usr/lib/nagios/plugins/check_zookeeper.py -s
"localhost:2181,localhost:2182,localhost:2183" -o nagios -k '$ARG1$' -w
'$ARG2$' -c '$ARG3$'
+ # ATTENTION: you should update the list of servers defined above
+}
+
+# 'check_zk_node' command definition
+define command {
+ command_name check_zk_node
+ command_line /usr/lib/nagios/plugins/check_zookeeper.py -s
$HOSTADDRESS$:2181 -o nagios -k '$ARG1$' -w '$ARG2$' -c '$ARG3$'
+ # ATTENTION: you should update the port. default: 2181
+}
+
Added: hadoop/zookeeper/trunk/src/contrib/monitoring/test.py
URL:
http://svn.apache.org/viewvc/hadoop/zookeeper/trunk/src/contrib/monitoring/test.py?rev=963957&view=auto
==============================================================================
--- hadoop/zookeeper/trunk/src/contrib/monitoring/test.py (added)
+++ hadoop/zookeeper/trunk/src/contrib/monitoring/test.py Wed Jul 14 06:39:32
2010
@@ -0,0 +1,267 @@
+#! /usr/bin/env python
+
+import unittest
+import socket
+import sys
+
+from StringIO import StringIO
+
+from check_zookeeper import ZooKeeperServer, NagiosHandler, CactiHandler,
GangliaHandler
+
+ZK_MNTR_OUTPUT = """zk_version\t3.4.0--1, built on 06/19/2010 15:07 GMT
+zk_avg_latency\t1
+zk_max_latency\t132
+zk_min_latency\t0
+zk_packets_received\t640
+zk_packets_sent\t639
+zk_outstanding_requests\t0
+zk_server_state\tfollower
+zk_znode_count\t4
+zk_watch_count\t0
+zk_ephemerals_count\t0
+zk_approximate_data_size\t27
+zk_open_file_descriptor_count\t22
+zk_max_file_descriptor_count\t1024
+"""
+
+ZK_MNTR_OUTPUT_WITH_BROKEN_LINES = """zk_version\t3.4.0
+zk_avg_latency\t23
+broken-line
+
+"""
+
+ZK_STAT_OUTPUT = """Zookeeper version: 3.3.0-943314, built on 05/11/2010 22:20
GMT
+Clients:
+ /0:0:0:0:0:0:0:1:34564[0](queued=0,recved=1,sent=0)
+
+Latency min/avg/max: 0/40/121
+Received: 11
+Sent: 10
+Outstanding: 0
+Zxid: 0x700000003
+Mode: follower
+Node count: 4
+"""
+
+class SocketMock(object):
+ def __init__(self):
+ self.sent = []
+
+ def settimeout(self, timeout):
+ self.timeout = timeout
+
+ def connect(self, address):
+ self.address = address
+
+ def send(self, data):
+ self.sent.append(data)
+ return len(data)
+
+ def recv(self, size):
+ return ZK_MNTR_OUTPUT[:size]
+
+ def close(self): pass
+
+class ZK33xSocketMock(SocketMock):
+ def __init__(self):
+ SocketMock.__init__(self)
+ self.got_stat_cmd = False
+
+ def recv(self, size):
+ if 'stat' in self.sent:
+ return ZK_STAT_OUTPUT[:size]
+ else:
+ return ''
+
+class UnableToConnectSocketMock(SocketMock):
+ def connect(self, _):
+ raise socket.error('[Errno 111] Connection refused')
+
+def create_server_mock(socket_class):
+ class ZooKeeperServerMock(ZooKeeperServer):
+ def _create_socket(self):
+ return socket_class()
+ return ZooKeeperServerMock()
+
+class TestCheckZookeeper(unittest.TestCase):
+
+ def setUp(self):
+ self.zk = ZooKeeperServer()
+
+ def test_parse_valid_line(self):
+ key, value = self.zk._parse_line('something\t5')
+
+ self.assertEqual(key, 'something')
+ self.assertEqual(value, 5)
+
+ def test_parse_line_raises_exception_on_invalid_output(self):
+ invalid_lines = ['something', '', 'a\tb\tc', '\t1']
+ for line in invalid_lines:
+ self.assertRaises(ValueError, self.zk._parse_line, line)
+
+ def test_parser_on_valid_output(self):
+ data = self.zk._parse(ZK_MNTR_OUTPUT)
+
+ self.assertEqual(len(data), 14)
+ self.assertEqual(data['zk_znode_count'], 4)
+
+ def test_parse_should_ignore_invalid_lines(self):
+ data = self.zk._parse(ZK_MNTR_OUTPUT_WITH_BROKEN_LINES)
+
+ self.assertEqual(len(data), 2)
+
+ def test_parse_stat_valid_output(self):
+ data = self.zk._parse_stat(ZK_STAT_OUTPUT)
+
+ result = {
+ 'zk_version' : '3.3.0-943314, built on 05/11/2010 22:20 GMT',
+ 'zk_min_latency' : 0,
+ 'zk_avg_latency' : 40,
+ 'zk_max_latency' : 121,
+ 'zk_packets_received': 11,
+ 'zk_packets_sent': 10,
+ 'zk_server_state': 'follower',
+ 'zk_znode_count': 4
+ }
+ for k, v in result.iteritems():
+ self.assertEqual(v, data[k])
+
+ def test_recv_valid_output(self):
+ zk = create_server_mock(SocketMock)
+
+ data = zk.get_stats()
+ self.assertEqual(len(data), 14)
+ self.assertEqual(data['zk_znode_count'], 4)
+
+ def test_socket_unable_to_connect(self):
+ zk = create_server_mock(UnableToConnectSocketMock)
+
+ self.assertRaises(socket.error, zk.get_stats)
+
+ def test_use_stat_cmd_if_mntr_is_not_available(self):
+ zk = create_server_mock(ZK33xSocketMock)
+
+ data = zk.get_stats()
+ self.assertEqual(data['zk_version'], '3.3.0-943314, built on
05/11/2010 22:20 GMT')
+
+class HandlerTestCase(unittest.TestCase):
+
+ def setUp(self):
+ try:
+ sys._stdout
+ except:
+ sys._stdout = sys.stdout
+
+ sys.stdout = StringIO()
+
+ def tearDown(self):
+ sys.stdout = sys._stdout
+
+ def output(self):
+ sys.stdout.seek(0)
+ return sys.stdout.read()
+
+
+class TestNagiosHandler(HandlerTestCase):
+
+ def _analyze(self, w, c, k, stats):
+ class Opts(object):
+ warning = w
+ critical = c
+ key = k
+
+ return NagiosHandler().analyze(Opts(), {'localhost:2181':stats})
+
+ def test_ok_status(self):
+ r = self._analyze(10, 20, 'a', {'a': 5})
+
+ self.assertEqual(r, 0)
+ self.assertEqual(self.output(), 'Ok "a"!|localhost:2181=5;10;20\n')
+
+ r = self._analyze(20, 10, 'a', {'a': 30})
+ self.assertEqual(r, 0)
+
+ def test_warning_status(self):
+ r = self._analyze(10, 20, 'a', {'a': 15})
+ self.assertEqual(r, 1)
+ self.assertEqual(self.output(),
+ 'Warning "a" localhost:2181!|localhost:2181=15;10;20\n')
+
+ r = self._analyze(20, 10, 'a', {'a': 15})
+ self.assertEqual(r, 1)
+
+ def test_critical_status(self):
+ r = self._analyze(10, 20, 'a', {'a': 30})
+ self.assertEqual(r, 2)
+ self.assertEqual(self.output(),
+ 'Critical "a" localhost:2181!|localhost:2181=30;10;20\n')
+
+ r = self._analyze(20, 10, 'a', {'a': 5})
+ self.assertEqual(r, 2)
+
+ def test_check_a_specific_key_on_all_hosts(self):
+ class Opts(object):
+ warning = 10
+ critical = 20
+ key = 'latency'
+
+ r = NagiosHandler().analyze(Opts(), {
+ 's1:2181': {'latency': 5},
+ 's2:2181': {'latency': 15},
+ 's3:2181': {'latency': 35},
+ })
+ self.assertEqual(r, 2)
+ self.assertEqual(self.output(),
+ 'Critical "latency" s3:2181!|s1:2181=5;10;20 '\
+ 's3:2181=35;10;20 s2:2181=15;10;20\n')
+
+class TestCactiHandler(HandlerTestCase):
+ class Opts(object):
+ key = 'a'
+ leader = False
+
+ def __init__(self, leader=False):
+ self.leader = leader
+
+ def test_output_values_for_all_hosts(self):
+ r = CactiHandler().analyze(TestCactiHandler.Opts(), {
+ 's1:2181':{'a':1},
+ 's2:2181':{'a':2, 'b':3}
+ })
+ self.assertEqual(r, None)
+ self.assertEqual(self.output(), 's1_2181:1 s2_2181:2')
+
+ def test_output_single_value_for_leader(self):
+ r = CactiHandler().analyze(TestCactiHandler.Opts(leader=True), {
+ 's1:2181': {'a':1, 'zk_server_state': 'leader'},
+ 's2:2181': {'a':2}
+ })
+ self.assertEqual(r, 0)
+ self.assertEqual(self.output(), '1\n')
+
+
+class TestGangliaHandler(unittest.TestCase):
+
+ class TestableGangliaHandler(GangliaHandler):
+ def __init__(self):
+ GangliaHandler.__init__(self)
+ self.cli_calls = []
+
+ def call(self, cli):
+ self.cli_calls.append(' '.join(cli))
+
+ def test_send_single_metric(self):
+ class Opts(object):
+ @property
+ def gmetric(self): return '/usr/bin/gmetric'
+ opts = Opts()
+
+ h = TestGangliaHandler.TestableGangliaHandler()
+ h.analyze(opts, {'localhost:2181':{'latency':10}})
+
+ cmd = "%s -n latency -v 10 -t uint32" % opts.gmetric
+ assert cmd in h.cli_calls
+
+if __name__ == '__main__':
+ unittest.main()
+
Propchange: hadoop/zookeeper/trunk/src/contrib/monitoring/test.py
------------------------------------------------------------------------------
svn:executable = *
