20 February 2013 Karim Elatov

In my home environment I only have 3 machines. I wanted to monitor them just for my own sake. At my previous job we used Nagios for such a task, but we also had thousands of machines and services to monitor. At home that is not the case. I just wanted a simple setup to show me CPU, memory, and network usage and that is about it.

*nix Systems

My FreeBSD machine:

freebsd:~>uname -smr
FreeBSD 9.1-RELEASE i386

My Fedora Machine:

moxz:~>uname -sr
Linux 3.7.8-202.fc18.i686
moxz:~>lsb_release -rdc
Description: Fedora release 18 (Spherical Cow)
Release: 18
Codename: SphericalCow

My Ubuntu Machine:

kerch:~>uname -sr
Linux 3.2.0-23-powerpc-smp
kerch:~>lsb_release -drc
Description: Ubuntu 12.10
Release: 12.10
Codename: quantal

Monitoring Systems

As I kept researching, I discovered that there are many different monitoring applications out there. Most are included in the wikipedia page “Comparison of network monitoring systems”. Checking out other sites, I saw many different comparisons:

As I mentioned, I have used Nagios before so I wanted to try something new. We also used Cacti, alongside with Nagios so I didn’t want to use that either. I cared about two aspects: simplicity and performance. Having said that, I decided to try out Collectd, from this site:

Collectd is similar to Munin and Cacti in that it focuses on graphing system metrics. Where it excels in is that it is designed specifically for performance and portability; this ultimately means it’s great on rugged systems, low-end systems, and embedded systems. Being designed for performance and low-system resource use means that Collectd can gather data every 10 seconds without interfering with your server processes, providing extremely high-resolution statistics.

Then I wanted to try out Monitorix, from their own site:

Monitorix is a free, open source, lightweight system monitoring tool designed to monitor as many services and system resources as possible. It has been created to be used under production Linux/UNIX servers, but due to its simplicity and small size can be used on embedded devices as well.

But then checking out their configuration page, I saw this for their network setup:

REMOTEHOST_LIST This is a list of the remote servers where Monitorix it's already installed and working and you plan to monitor them from this one. It consists of a pair of values being in the left side the description of each server and in the right side the URL or IP address. An example of this list would be:

our @REMOTEHOST_LIST = ( "WWW Linux", "http://www.example.com", "Backup Linux", "http://192.168.1.4", "SMTP Linux", "http://71.16.11.2:8080", );

As you can see all three entries use URLs to designate the location of each remote server. This means that on each server most also have been installed a CGI capable web server like Apache.

I didn’t want to run a webserver on each of my clients just so I could monitor them. So I decided to skip Monitorix.

I also wanted to try out Munin, from this site:

One of Munin’s greatest strengths is how simple it is to extend. With just a few lines of code, you can write a plugin to monitor almost anything. Being so easy to extend means that Munin is also a good choice for graphing things unrelated to server performance, such as the number of user signups or website popularity.

Also from here:

The primary emphasis of Munin is on the plug and play architecture for it’s plugin. There are lot of plugins available for Munin, which will just work out-of-the box without lot of tweaking.

Lastly I wanted to compare it to a large application just to see it’s Pro’s and Con’s. Since Nagios and Cacti were out of the picture, I decided to try Zenoss. There are a couple of sites that talk about the differences between Nagios and Zenoss, here are a few:

Zenoss seemed comparable :)

Since I picked 3 different applications, I will break this post into 3 different parts; one per application.

Collectd

To send information across the network we need to configure a collector (server) and nodes (clients). Instructions on how to configure Collectd for such a setup are here: “Networking introduction

1. Install Collectd On Ubuntu and Configure it as a Collector/Server

First let’s install the software:

kerch:~>sudo apt-get install collectd

After the install finishes, configure it. To do so, edit the /etc/collectd/collectd.conf file and enable any plugins you desire. Here is how my setup looked like:

kerch:~>grep -v -E '^$|^#' /etc/collectd/collectd.conf
FQDNLookup true
LoadPlugin syslog
<plugin syslog>
  LogLevel info
</plugin>
LoadPlugin cpu
LoadPlugin disk
LoadPlugin interface
LoadPlugin load
LoadPlugin memory
LoadPlugin network
LoadPlugin nfs
LoadPlugin rrdtool
LoadPlugin swap
<plugin network>
  Listen "192.168.1.100"
  <listen "192.168.1.100" >
    SecurityLevel Sign
    AuthFile "/etc/collectd/passwd"
    Interface "eth0"
  </listen>
  MaxPacketSize 1452
  CacheFlush 1800
</plugin>
<plugin rrdtool>
  DataDir "/var/lib/collectd/rrd"
</plugin>
Include "/etc/collectd/filters.conf"
Include "/etc/collectd/thresholds.conf"

The SecurityLevel setup is probably unnecessary, but I wanted to try it out just in case. Here is how my password file looked like:

kerch:~>cat /etc/collectd/passwd
elatov:test

I was actually running iptables on my Ubuntu machine, so I had to allow nodes/clients to connect to my Collectd collector. To do that, I edited the /etc/iptables/rules.v4 file and added the following to it:

-A INPUT -s 192.168.1.0/24 -p udp -m state --state NEW -m udp --dport 25826 -j ACCEPT

I then restarted my iptables instance to apply the changes:

kerch:~>sudo service iptables-persistent restart
* Loading iptables rules...
* IPv4...
* IPv6... [ OK ]

Then I started the collectd Service:

kerch:~>sudo service collectd start
Starting statistics collection and monitoring daemon: collectd.

2. Install Collectd on FreeBSD and set it up as Client

Let’s find the software:

freebsd:~>whereis collectd
collectd: /usr/ports/net-mgmt/collectd

Now let’s go ahead and install it:

freebsd:~>cd /usr/ports/net-mgmt/collectd
freebsd:/usr/ports/net-mgmt/collectd>sudo make install clean

At that point the compile process will fire up and install the software. Just for reference here are the configurations for the collectd package and it’s prerequisites:

freebsd:/usr/ports/net-mgmt/collectd>make showconfig
===> The following configuration options are available for collectd-4.10.8_3:
     BIND=off: Enable BIND 9.5+ statistics
     CGI=on: Install collection.cgi (requires RRDTOOL)
     DEBUG=off: Enable debugging
     GCRYPT=on: Build with libgcrypt
     VIRT=off: Build with libvirt
====> Options available for the group INPUT
     APACHE=off: Apache mod_status (libcurl)
     APCUPS=off: APC UPS (apcupsd)
     CURL=off: CURL generic web statistics
     CURL_JSON=off: CURL JSON generic web statistics
     CURL_XML=off: CURL XML generic web statistics
     DBI=off: database abstraction library
     DISK=on: Disk performance statistics
     NUTUPS=off: NUT UPS daemon
     INTERFACE=on: Network interfaces (libstatgrab)
     MBMON=off: MBMon
     MEMCACHED=off: Memcached
     MYSQL=off: MySQL
     NGINX=off: Nginx
     OPENVPN=off: OpenVPN statistics
     PDNS=off: PowerDNS
     PGSQL=off: PostgreSQL
     PING=on: Network latency (liboping)
     PYTHON=off: Python plugin
     ROUTEROS=off: RouterOS plugin
     SNMP=on: SNMP
     TOKYOTYRANT=off: Tokyotyrant database
     XMMS=off: XMMS
====> Options available for the group OUTPUT
     RRDTOOL=on: RRDTool
     RRDCACHED=on: RRDTool Cached (require RRDTOOL)
     WRITE_HTTP=off: write_http
===> Use 'make config' to modify these settings

Here is the config for rrdtool:

freebsd:/usr/ports/databases/rrdtool>make showconfig
===> The following configuration options are available for rrdtool-1.4.7_2:
     DEJAVU=off: Use DejaVu fonts (requires X11)
     JSON=off: Support of json export
     MMAP=on: Use mmap in rrd_update
     PERL_MODULE=on: Build PERL module
     PYTHON_MODULE=off: Build PYTHON bindings
     RUBY_MODULE=off: Build RUBY bindings
===> Use 'make config' to modify these settings

Lastly here is the configuration for net-snmp:

freebsd:/usr/ports/net-mgmt/net-snmp>make showconfig
===> The following configuration options are available for net-snmp-5.7.2_2:
     AX_SOCKONLY=off: Disable UDP/TCP transports for agentx
     DMALLOC=off: Enable dmalloc debug memory allocator
     DUMMY=on: Enable dummy values as placeholders
     IPV6=off: IPv6 protocol
     MFD_REWRITES=off: Build with 64-bit Interface Counters
     MYSQL=off: MySQL database
     PERL=on: Perl scripting language
     PERL_EMBEDDED=on: Build embedded perl
     PYTHON=off: Python bindings
     TKMIB=off: Install graphical MIB browser
     UNPRIVILEGED=off: Allow unprivileged users to execute net-snmp
===> Use 'make config' to modify these settings

After the software is installed, we need to configure it. Edit the /usr/local/etc/collectd.conf file and make the necessary changes. Here is how my file looked like:

freebsd:~>grep -v -E '^$|^#' /usr/local/etc/collectd.conf
Hostname "freebsd.dnsd.me"
FQDNLookup true
LoadPlugin syslog
<plugin syslog>
  LogLevel info
</plugin>
LoadPlugin cpu
LoadPlugin disk
LoadPlugin interface
LoadPlugin load
LoadPlugin memory
LoadPlugin network
LoadPlugin rrdtool
LoadPlugin swap
<plugin network>
  # client setup:
  Server "192.168.1.100"
  <server "192.168.1.100">
    SecurityLevel Encrypt
    Username "elatov"
    Password "test"
  </server>
  CacheFlush 1800
</plugin>
<plugin rrdtool>
  DataDir "/var/lib/collectd/rrd"
  CacheTimeout 120
  CacheFlush 900
</plugin>

Now let’s enable the daemon, this is done by editing /etc/rc.conf and adding the following:

collectd_enable="YES"

Now let’s start the collectd daemon:

freebsd:~>sudo /usr/local/etc/rc.d/collectd start

At this point we can check to make sure the files are now getting uploaded to the collector machine (our Ubuntu Machine):

kerch:~>ls -1 /var/lib/collectd/rrd/
freebsd.dnsd.me
kerch.dnsd.me

That looks good, now let’s install a web GUI for Collectd.

3. Install Collectd-Web Front End on the Collector

There are multiple options, this page has a list of available front ends for Collectd.

I decided to go with Collectd-Web cause it seemed simple. Instructions on how to install the software can be found here. First download the necessary files:

kerch:~>git clone git://github.com/httpdss/collectd-web.git

then make sure all the dependencies look good:

kerch:~>cd collectd-web
kerch:~/collectd-web>./check_deps.sh
Carp looks ok
CGI looks ok
CGI::Carp looks ok
HTML::Entities looks ok
URI::Escape looks ok
RRDs looks ok
Data::Dumper looks ok
JSON looks ok

Now let’s just copy the whole directory to a folder where your web Server (I was using Apache for other uses) can host the files:

kerch:~>sudo rsync -avzP collectd-web/. /var/www/cw/.

Lastly make sure collectd looks under the correct location (where the RRD files are stored):

kerch:~>cat /etc/collectd/collection.conf
datadir: "/var/lib/collectd/rrd/"
libdir: "/usr/lib/collectd/"

This should be the default, but just in case. Now visiting the Collectd-Web portal (http://localhost/cw), I saw the following:

collectd web first page Monitor Different Systems with Collectd

Selecting the remote host (freebsd) and then the CPU option, allowed me to see the CPU stats of my FreeBSD machine:

collectd web freebsd cpu Monitor Different Systems with Collectd

Now let’s configure the Fedora machine to send information to the collector.

4. Install Collectd on Fedora and Configure it as a Client

First install the collectd package:

moxz:~>sudo yum install collectd collectd-rrdtool

Then I edited the /etc/collectd.conf and configured the same things as before:

moxz:~>grep -v -E '^#|^$' /etc/collectd.conf
Hostname "moxz.dnsd.me"
FQDNLookup true
LoadPlugin syslog
<plugin syslog>
LogLevel info
</plugin>
LoadPlugin cpu
LoadPlugin disk
LoadPlugin interface
LoadPlugin load
LoadPlugin memory
LoadPlugin network
LoadPlugin nfs
LoadPlugin rrdtool
<plugin network>
  Server 192.168.1.100
  <server "192.168.1.100">
    SecurityLevel Encrypt
    Username "elatov"
    Password "test"
    Interface "eth0"
  </server>
  CacheFlush 1800
</plugin>
Include "/etc/collectd.d"
<plugin rrdtool>
  DataDir "/var/lib/collectd/rrd"
  CacheTimeout 120
  CacheFlush 900
</plugin>

I then enabled the service:

moxz:~>sudo systemctl enable collectd
ln -s '/usr/lib/systemd/system/collectd.service' '/etc/systemd/system/multi-user.target.wants/collectd.service'

Starting the service looked like this:

moxz:~>sudo systemctl start collectd
moxz:~>sudo systemctl status collectd
collectd.service - Collectd
          Loaded: loaded (/usr/lib/systemd/system/collectd.service; enabled)
          Active: active (running) since Wed 2013-02-27 13:08:06 PST; 2 days ago
        Main PID: 711 (collectd)
          CGroup: name=systemd:/system/collectd.service
                  └─711 /usr/sbin/collectd -C /etc/collectd.conf -f

I then visited the same Collectd-Web portal and saw the new host’s stats.

All of the above was for Collectd version 4. A newer version is out, version 5, and migration steps are available here. Make sure at least the collector is at version 5, but it would be best to have both collector and nodes be at the same version.

The last thing I wanted to do was check on the status of my raid on the FreeBSD machine.

5. Configure a Custom Script On the FreeBSD Collectd Client

There is a command line program called arcconf which allows you to see the status. Here is how it looks:

freebsd:~>sudo arcconf getconfig 1
Controllers found: 1
----------------------------------------------------------------------
Controller information
----------------------------------------------------------------------
   Controller Status                        : Optimal
   Channel description                      : SATA
   Controller Model                         : CERC SATA1.5/6ch
   Controller Serial Number                 : BBE5AA
   Installed memory                         : 64 MB
   Copyback                                 : Disabled
   Background consistency check             : Enabled
   Automatic Failover                       : Enabled
   Stayawake period                         : Disabled
   Spinup limit internal drives             : 0
   Spinup limit external drives             : 0
   Defunct disk drive count                 : 0
   Logical devices/Failed/Degraded          : 1/0/0
   --------------------------------------------------------
   Controller Version Information
   --------------------------------------------------------
   BIOS                                     : 4.1-0 (7417)
   Firmware                                 : 4.1-0 (7417)
   Driver                                   : 2.1-9 (1)
   Boot Flash                               : 0.0-0 (0)
   --------------------------------------------------------
   Controller Battery Information
   --------------------------------------------------------
   Status                                   : Not Installed

----------------------------------------------------------------------
Logical device information
----------------------------------------------------------------------
Logical device number 0
   Logical device name                      : DATA 1
   RAID level                               : 1
   Status of logical device                 : Optimal
   Size                                     : 152554 MB
   Read-cache mode                          : Enabled
   Write-cache mode                         : Disabled (write-through)
   Write-cache setting                      : Disabled (write-through)
   Partitioned                              : Yes
   Protected by Hot-Spare                   : No
   Bootable                                 : Yes
   Failed stripes                           : No
   Power settings                           : Disabled
   --------------------------------------------------------
   Logical device segment information
   --------------------------------------------------------
   Segment 0                                : Present (Controller:1,Channel:0,Device:1) Y450QB3E
   Segment 1                                : Present (Controller:1,Channel:0,Device:0) Y450QA0E


----------------------------------------------------------------------
Physical Device information
----------------------------------------------------------------------
   Channel #0:
      Transfer Speed                        : SATA 1.5 Gb/s
      Device #0
         Device is a Hard drive
         State                              : Online
         Supported                          : Yes
         Transfer Speed                     : SATA 1.5 Gb/s
         Reported Channel,Device(T:L)       : 0,0(0:0)
         Vendor                             : Maxtor
         Model                              : 6Y160M0
         Firmware                           : YAR5
         Serial number                      : Y450QA0E
         Size                               : 152587 MB
         Write Cache                        : Unknown
         FRU                                : None
         S.M.A.R.T.                         : No
         S.M.A.R.T. warnings                : 0
         NCQ status                         : Disabled
      Device #1
         Device is a Hard drive
         State                              : Online
         Supported                          : Yes
         Transfer Speed                     : SATA 1.5 Gb/s
         Reported Channel,Device(T:L)       : 0,1(1:0)
         Vendor                             : Maxtor
         Model                              : 6Y160M0
         Firmware                           : YAR5
         Serial number                      : Y450QB3E
         Size                               : 152587 MB
         Write Cache                        : Unknown
         FRU                                : None
         S.M.A.R.T.                         : No
         S.M.A.R.T. warnings                : 0
         NCQ status                         : Disabled
Command completed successfully.

There are no SMART capabilities, so the only thing I can check is whether the disks are online. Here is a concise view:

freebsd:~>arcconf getconfig 1 PD| grep State
State : Online
State : Online

we have two disks that are online. Collectd has a plugin called exec, it allows you to run a command and plot values from the results of that command. More information can be seen this post. But I will settle with just plotting how many disks are currently online. So I wrote this script:

#!/usr/local/bin/bash
HOST=$(hostname -f)
INTERVAL=300
while sleep "$INTERVAL"; do
        val=$(/usr/local/sbin/arcconf getconfig 1 PD | grep State | grep Online| /usr/bin/wc -l)
        echo "PUTVAL \"$HOST/exec-raid/gauge\" interval=$INTERVAL N:$(eval echo \$val)"
done

Nothing fancy, it just counts how many disks are online and sets the interval to be 300 (5 minutes). As a quick test change the interval value to 10 and run the script, it should produce something similar to this:

freebsd:~>./arc.sh
PUTVAL "freebsd.dnsd.me/exec-raid/gauge" interval=10 N:2
PUTVAL "freebsd.dnsd.me/exec-raid/gauge" interval=10 N:2

Now let’s enable this to be executed from Collectd. Edit /usr/local/etc/collectd.conf and add/modify the following

<plugin exec>
        Exec "elatov:elatov" "/home/elatov/arc.sh"
</plugin>

After some time the graph started to populate with data. Here is a very small sample of how it looked like:

collectd web freebsd raid gauge Monitor Different Systems with Collectd


blog comments powered by Disqus