Educating the world

Our blog has over 10,000 readers a month

Installing OpenMeeting with MySQL

March 14th, 2013

From Wikipedia OpenMeeting is:

OpenMeetings is software used for presenting, online training, web conferencing, collaborative whiteboard drawing and document editing, and user desktop sharing. The product is based on OpenLaszlo RIA framework and Red5 media server, which in turn are based on a bunch of open source components. Communication takes place in meeting rooms which are set to different communication, security and video quality modes. The recommended database for backend support is MySQL. The product can be set up as an installed server product, or used as a hosted product.

Today we will be installing OpenMeeting on CentOS 6.3 (64-bit).
Find a nice place to work:

mkdir /home/dev
cd /home/dev

Go to and download and unpack latest binary.

  1. wget

  2. mkdir openmeeting

  3. cd openmeeting

  4. tar -xzvf ../apache-openmeetings-incubating-2.0.0.r1361497-14-07-2012_1108.tar.gz

The default installation of OpenMeeting uses an integrated Apache Derby database to persist data at the back end. They recommend MySQL (or a real database) for production installations. I however, like to easily poke around the database and investigate how the product works. The trouble with an integrated database is that it won’t be available when the application is down and it may not be available to clients outside the application.

OpenMeeting requires UTF8 so we’ll install MySQL and configure the collation.

  1. Install PHP and MySQL with yum
  2. yum install php mysql-server mysql

  3. There are a few problems of setting the collation with MySQL so. Edit /etc/my.cnf and add the following lines to the mysqld section:

    collation-server = utf8_unicode_ci
    init-connect=’SET NAMES utf8′
    character-set-server = utf8

  4. Then restart MySQL with:

    /etc/init.d/mysqld restart

Now that we have MySQL installed, we’ll have to create a MySQL account for OpenMeeting to use.

  1. Login to MySQL:

    mysql -uroot

  2. Create the database:

    CREATE DATABASE openmeetings;

  3. Create a MySQL user for the application:

    CREATE USER openmeetings;

  4. Set permissions:

    GRANT ALL ON openmeetings.* TO openmeetings@localhost;

  5. Set password (change **** to your password):

    SET PASSWORD FOR openmeetings@localhost=PASSWORD(’****’);

  6. Log out of MySQL:


  7. Now we will just check that our new user is set up properly. (Where **** is the password)

    mysql -uopenmeetings -p**** openmeetings

OpenMeeting isn’t bundled with all the database connectors so we’ll have to download and install the MySQL connector in order to talk to the database.

  1. Start off in our dev area:

    cd /home/dev

  2. Go to and download the latest ConnectorJ (at the time of writing this was 5.1.24)
  3. Unpack it:

    tar -xvzf mysql-connector-java-5.1.24.tar.gz

  4. Move the connector to the correct place in the OpenMeeting directory structure so it can be found by the application:

    mv mysql-connector-java-5.1.24/mysql-connector-java-5.1.24-bin.jar openmeeting/webapps/openmeetings/WEB-INF/lib/

We are not using the default database type so we must change the configuration files so that the OpenMeeting installation process uses the MySQL configuration file instead of the integrated Derby configuration file.

  1. Go to the connectors folder:

    cd /home/dev/openmeeting/webapps/openmeetings/WEB-INF/classes/META-INF

  2. Use the MySQL configuration template:

    cp mysql_persistence.xml persistence.xml

  3. Edit persistence.xml and change the MySQL credentials. At the bottom of the file you will see

    , Url=jdbc:mysql://localhost:3306/openmeetings?autoReconnect….
    …. , MaxActive=100
    , MaxWait=10000
    , TestOnBorrow=true
    , poolPreparedStatements=true
    , Username=openmeetings
    , Password=****"/>

    Change the Username and Password to what you set earlier. If you want to change the database name localhost:3306/openmeetings? will become localhost:3306/my_database?.

Now we are going to set up Enabling Image Upload and import to whiteboard.

To install ImageMagick we’ll use the package manager:

yum install ImageMagick

Installing ImageMagick will install GhostScript as a dependency so we get part 1 of Enabling import of PDFs into whiteboard for free. Part 2 requires installing SWFTools which we will do now:

  1. Start in our development directory:

    cd /home/dev

  2. Go to and get the latest version:


  3. Unpack with:

    tar -xvzf swftools-0.9.2.tar.gz

  4. If this is a minimal version of CentOS (or you type gcc and get bad command) then you’ll need to install the C compiler and tools:

    yum install gcc* automake zlib-devel libjpeg-devel giflib-devel freetype-devel make

  5. Prepare to build the software:

    cd /home/dev/swftools-0.9.2

  6. Run the pre-build configuration script:


  7. Build SWFTools:


There is a bug in the make install step so we’ll just fix that before running it.

  1. Edit /home/dev/swftools-0.9.2/swfs/Makefile
  2. Search for the install: directive and change:

            rm -f $(pkgdatadir)/swfs/default_viewer.swf -o -L $(pkgdatadir)/swfs/default_viewer.swf
            $(LN_S) $(pkgdatadir)/swfs/simple_viewer.swf $(pkgdatadir)/swfs/default_viewer.swf
            rm -f $(pkgdatadir)/swfs/default_loader.swf -o -L $(pkgdatadir)/swfs/default_loader.swf
            $(LN_S) $(pkgdatadir)/swfs/tessel_loader.swf $(pkgdatadir)/swfs/default_loader.swf


    <TAB>rm -f $(pkgdatadir)/swfs/default_viewer.swf
    <TAB>$(LN_S) $(pkgdatadir)/swfs/simple_viewer.swf $(pkgdatadir)/swfs/default_viewer.swf
    <TAB>rm -f $(pkgdatadir)/swfs/default_loader.swf
    <TAB>$(LN_S) $(pkgdatadir)/swfs/tessel_loader.swf $(pkgdatadir)/swfs/default_loader.swf

    Make sure that the first and only character at the start of the replacement lines is a TAB and not a space.

  3. Now we can run:

    make install

Next we’ll install the dependencies for Enabling import of .doc, .docx, .ppt, .pptx, … all Office Documents into whitebaord. This also installs Oracle’s Java which is why we didn’t install it earlier.

  1. Install Libre Office:

    yum install libreoffice-writer

Enabling Recording and import of .avi, .flv, .mov and .mp4 into whiteboard requires FFMpeg which is not available via standard yum so we’ll add another repository that does contain it. Thanks to needlessly does a “yum update” which would probably update everything seeing as CentOS keep stable older version instead of the bleeding edge.

  1. Edit /etc/yum.repos.d/dag.repo
  2. Add:

    name=DAG RPM Repository

  3. Import keys so we can talk to the repository securely:

    rpm −−import

  4. Start installation:

    yum install ffmpeg

Enabling Recording and import of .avi, .flv, .mov and .mp4 into whiteboard also requires SOX but that is available on the standard yum so just install that.

yum install sox

I have read through the installation instructions for OpenMeeting and they are as clear as mud. Half the documents seem to say that JODConverter is needed and the other half say that it isn’t. Even the project page for JODConverter says that the project isn’t even maintained. It has some connection to OpenOffice/LibreOffice, so I’m installing it anyway just in case.

  1. Change back to our scratch folder:

    cd /home/dev

  2. Go to
  3. And download the latest version:


  4. Then unzip it


Well that’s it for the dependencies. Hopefully OpenMeeting will find all the things it needs during it installation.

After following the original instructions for installing OpenMeeting, the Oracle Java had been replaced by the GNU version of the Java runtime. When I started the OpenMeeting container it use that version instead of the Oracle implementation and didn’t work.
If you install the dependencies in the order that I described above then Oracle’s Java should come out on top. We need to check to make sure though:

  1. Get which version of Java are you running:

    java -version

  2. You should get:

    java version “1.7.0_13″
    Java™ SE Runtime Environment (build 1.7.0_13-b20)
    Java HotSpot™ 64-Bit Server VM (build 23.7-b01, mixed mode)

  3. If you don’t then you’ll have to uninstall the GNU version:

    yum remove gij
    rpm -e jre

  4. And reinstall Oracle’s Java over the top:

    1. Navigate to:
    2. And download the appropriate version I selected “Linux x64 RPM".
    3. Install via the Redhat Package Manager, you might see a couple of errors like “Error: Could not open input file” but you can ignore these.

      rpm -i jre-7u13-linux-x64.rpm

OpenMeeting uses a couple of ports to communicate with so make sure your firewall allows traffic on them. For development purposes the following will remove all firewall rules until you reboot.

iptables -F

To start OpenMeeting:

  1. Change directory:

    cd /home/dev/openmeeting

  2. Type:


It should print a lot of information on the screen. There will be pauses and gaps but when it starts repeating no Appointments in range you are ready for the next step.

You can check [crudely] that the database has been created in MySQL by ensuring that there are plenty of files in:

The final part of the installation is run from the web interface so navigate to:
where localhost is the machine you have installed OpenMeeting on.

Click the link Continue with STEP 1. This will take you to the initial configuration screen where you can set up the application:

In the Userdata section fill in the admin user

Username: admin
Password: admin

In the Configuration section I didn’t touch anything but the names seem pretty self explanatory if you know how to set up your mail system (which is beyond the scope of this document).

In the Converters section we need to add the paths to the support utilities as they won’t be available to the servlet container.

All the Converter support applications should be in your path except jodconverter-core. I still don’t understand if or how JODConverter is needed, so add a reference to the JOD path anyway.


I leave all the other stuff to their defaults.

Finally click INSTALL. The screen will spend a couple of minutes spewing out data, so let it get on with it. When it has finished your web browser will sent you to the “Installation Complete!” page.

Click Enter the Application and login with Username as admin and Password as admin.

The rest is up to you! Good luck.

Killing a crashed Task Manager

March 7th, 2013

I’m running Vista at the moment and it is terrible. I think we are all tired of the appalling file copying mechanism. I came across a new bug in Vista today and thought I’d blog how I got out of it.

Task manager was loaded in and minimized.


  1. Right-clicked on the Task manager icon in the systray and the menu appeared. The menu items were click-able but the clicks wouldn’t register. The menu wouldn’t disappear after that, no matter what you did.
  2. The minimized icon in the systray was spiking, It appeared to be flashing as it rapidly switched from high CPU to low CPU.
  3. I couldn’t start any explorer sessions using the following methods:

    1. The WindowsKey-E wouldn’t work.
    2. Running explorer from Start->Run
    3. Clicking New Task… inside the Task manager and running explorer


We must kill Task manager from the command line.

  1. First we need to start a Powershell session, because a normal command window doesn’t have the right tools in its path. Start->Run


  2. Next list the processes currently running on the machine.

    PS C:\Users\mrn> Get-Process
    Handles NPM(K) PM(K) WS(K) VM(M) CPU(s) Id ProcessName
    ——- —— —– —– —– —— – ———–

    118 9 4552 11972 98 0.95 4132 taskmgr

  3. Make a note of the the Id of the Task manager process (taskmgr).
  4. Powershell makers have created a “kill” alias to the Stop-Process command to make it look a bit more like wonderful unix.

    kill 4132

  5. You may find that the systray icon is still present. If you hoover over the icon it will disappear.

Configuring a permanant DHCP reservation on a Cisco ASA/PIX

February 2nd, 2013

One of our offices uses a Cisco ASA/PIX and we want to manage all the IP address allocations with DHCP. The main benefit of this is that the dynamic IP address allocations can be managed centrally. If we change the default gateway of the network then no one needs to make any changes to the network settings on their devices.

Each device on the network gets an IP address when they ask for it and keeps it for a fixed amount of time called a lease. When that time has expired the device releases the address and asks for another. In most cases the device will be given the same information again by the DHCP server.

If you have a network with a lot of laptops, phones, tablets or printers then devices will come and go quite frequently. You will find that if you switch off the device for any amount of time it will come back with a different address. For laptops and personal devices this doesn’t matter. However if the device is a shared resource like a printer or file server then it can be a problem. Anyone who still wants to use that shared resource now needs to know about its new address.

We want to be able to tell the DHCP server that while it can allocate IP addresses from a certain pool we want to make sure that it can only allocate some of them to specific network devices. This will ensure that if a printer goes for repair and comes back in a week, when it’s switched back on it will have the same IP address that it always had.

Unfortunately there isn’t a structured way of doing this with a Cisco ASA/PIX so we need to find a work-a-round.

The following instructions describe how to do this but they also describe how to get into the administration section of the Cisco ASA/PIX because most of the instructions (on the internet) assume you know how to do this already. There is a strong argument that you should know what you are doing before you play with a router/firewall’s configuration but if your network man is on holiday then you might have to get your hands dirty yourself.

There are various GUIs to help with this kind of administration and that’s fine if you are at the customer’s site but most of the time I’m not. Short of connecting to the VPN and using the management console the easiest way is to go in on the command line using telnet.

In the following scenario we have been told that the printer is set to use DHCP and its current IP address is We want to add that to the pool and make sure it is given the same IP address each time.

So connect to the Cisco ASA/PIX. There’s no user name only a password, so enter the user level password:

host# telnet
User Access Verification

Password: ***
Type help or ‘?’ for a list of available commands.

Once logged in we need to switch to the administration mode.

cisco> enable
Password: ***

The user told us the printer was currently switched on so we can read the router’s Address Resolution Protocol table which lists the mappings between IP addresses and Media Access Control address (MAC address or address network card address). We’ll need the MAC address as it is the reference the router talks to whereas the IP address is only an abstraction.

cisco# show arp
inside 0c1b.ae43.bd21

Now we can check that the address is in the pool of DHCP addresses

cisco# show running-config dhcpd
dhcpd address inside
dhcpd dns interface inside
dhcpd domain interface inside
dhcpd enable inside

which it is and we can check that the MAC address is not currently assigned to anything else.

cisco# show running-config arp

Our sanity checks are ok so we are ready to proceed with the update. We must enter the configuration section by specifying that we will change the configuration from the terminal.

cisco# configure terminal

Once in the configuration section we can start changing the settings. The following line says that when we see the MAC address 0c1b.ae43.bd21 we are going to statically refer to it with the IP address of The Cisco ASA/PIX knows that this is already in the DHCP pool and won’t allocate it again.

cisco(config)# arp inside 0c1b.ae43.bd21

We can check the change has been added with the following command:

cisco(config)# show running-config arp
arp inside 0c1b.ae43.bd21

If you have miss-typed or you would like to remove an old entry you can do so by prefixing the existing command with no. For example:

no arp inside 0c1b.ae43.bd21

The changes are currently only made in memory, so we need to write the current running configuration down to disk.

cisco(config)# write mem
Building configuration…
Cryptochecksum: 389f1812 7c29dd7b 50ad4ca0 4ce3fd5e

4396 bytes copied in 1.480 secs (4396 bytes/sec)

And finally the job is done so we exit cleanly

cisco(config)# exit
cisco# exit


Connection closed by foreign host.

Rebooting the printer will result in the printer coming back with the same IP address.

Many thanks to goldplated for his original article.

Installing OpenVPN Client on MacOs Mountain Lion

January 18th, 2013

There are several OpenVPN clients for MacOS but the most popular one is Tunnelblick. At the time of writing the latest stable version is 3.2.8 but there is a red health warning saying that it’s not recommended for OS X 10.8 ("Mountain Lion"), where as the previous stable version 3.3beta21b version is. Yeh, weird that!

Version 3.3beta21b had one problem with launching the VPN tunnel, it couldn’t find the configurations specified in the .opvn configuration file. First we’all go through how to install it and then how to get around the bug with launching it.

  1. Your VPN administrator will have given you a zip file containing your OpenVPN configuration. It will have a name like
  2. Finder will automatically unpack it and create a folder called
  3. Go to Tunnelblick’s download page:
    and download Tunnelblick 3.3beta21b.
  4. Launch the download.
  5. Double-click the icon.
  6. Click I have configuration files.
  7. Click OpenVPN Configuration(s).
  8. Click Open Private Configuration Folder.
  9. The Finder will open.
  10. Drag the youname-12345 configuration folder to the same place where you see the Launch Tunnelblick icon.
  11. Click Done.
  12. Click Do not check for a change.
  13. Click Don’t check for automatic updates. The latest version is not compatible with Mountain Lion so we don’t want your Mac to automatically install it!

If you go up to the greyed out tunnel icon next to the time on the menu bar and click it you can select yourname-12345->Connect <VPN name>. It will pop up a message saying:

Tunnelblick was unable to start OpenVPN to connect yourname-12345/<VPN name>. For details, see the OpenVPN log in the VPN Details… window

Click the greyed out tunnel icon again and select VPN Details. For the sake of those people googling (searching) for a solution here is the error message from the log to bring you in!

2013-01-18 00:49:41 *Tunnelblick: OS X 10.6.8; Tunnelblick 3.3beta21b (build 3114.3185)
2013-01-18 00:49:41 *Tunnelblick: Attempting connection with yourname-12345/vpn-name; Set nameserver = 1; monitoring connection
2013-01-18 00:49:41 *Tunnelblick: /Applications/ start yourname-12345/vpn-name.ovpn 1337 1 0 0 0 49 -atADGNWradsgnw
2013-01-18 00:49:42 *Tunnelblick:

Could not start OpenVPN (openvpnstart returned with status #242)

Contents of the openvpnstart log:

OpenVPN returned with status 1, errno = 2:
No such file or directory

Command used to start OpenVPN (one argument per displayed line):

/Users/username/Library/Application Support/Tunnelblick/Configurations
/Users/username/Library/Application Support/Tunnelblick/Configurations/yourname-12345/vpn-name.ovpn
/Library/Application Support/Tunnelblick/Logs/-SUsers-Syourname-SLibrary-SApplication Support-STunnelblick-SConfigurations-Syourname-12345-Sclient–vpn-name.ovpn.1_0_0_0_49.1337.openvpn.log
/Applications/ -m -w -d -atADGNWradsgnw
/Applications/ -m -w -d -atADGNWradsgnw
/Applications/ -m -w -d -atADGNWradsgnw

Contents of the OpenVPN log:

Options error: –ca fails with ‘ca.crt’: No such file or directory
Options error: –cert fails with ‘yourname-12345.crt’: No such file or directory
Options error: –key fails with ‘yourname-12345.key’: No such file or directory
Options error: Please correct these errors.
Use –help for more information.

More details may be in the Console Log’s “All Messages”

Now the fix.

  1. Open a Finder window and go to:
    /Users/username/Library/Application Support/Tunnelblick/Configurations/yourname-12345
  2. Edit the file vpn-name.ovpn
  3. At the bottom of the file you will see the following lines:

    ca ca.crt
    key yourname-12345.key
    cert yourname-12345.crt

  4. The launcher was complaining that it couldn’t find the files so there is something wrong with how the software is working out which directory to look in.
  5. You must change the lines to include the full path to the file.

    ca “/Users/username/Library/Application Support/Tunnelblick/Configurations/yourname-12345/ca.crt”
    key “/Users/username/Library/Application Support/Tunnelblick/Configurations/yourname-12345/yourname-12345.key”
    cert “/Users/username/Library/Application Support/Tunnelblick/Configurations/yourname-12345/yourname-12345.crt”

  6. After the change, it doesn’t matter what the application thinks is the configuration directory as we are using the absolute path to the files instead of a relative one.
  7. Go up to the greyed out tunnel icon and select
    yourname-12345->Connect <VPN name>.
  8. And hay presto, you’re in.

Cloning a VirtualBox machine running CentOS causes network interfaces to fail

January 11th, 2013

When you clone a VirtualBox machine in order to create a new virtual machine (as opposed to creating a back up of an existing machine) then you must check Reinitialize the MAC address of all network cards and use the Full clone type to make sure that the original and the clone don’t share any resources.

After booting the new (freshly) cloned machine you will notice that all the network interfaces are down. Trying to raise them gives you the following:

[root@myhost ~]# ifup eth0
Device eth0 does not seem to be present, delaying initialization.
[root@myhost ~]# ifup eth1
/sbin/ifup: configuration for eth1 not found.
Usage: ifup <device name>

When Linux boots the clone image for the first time it sees the new MAC address. It compares it with the MAC address it has saved for eth0 and realises that they are different. In order to avoid a conflict (in case the old network card comes back), Linux creates a new interface called eth1.

We are running VirtualBox (2.4.6) and don’t have network cards in the virtual machine. We just want it to be the same as the original machine without a lot of faff.

  1. Edit the NIC cache:

    vi /etc/udev/rules.d/70-persistent-net.rules

  2. Delete the line where NAME="eth0″
  3. Find the line with NAME="eth1″ and change eth1 to eth0
  4. Save file
  5. Edit the new interface file:

    vi /etc/sysconfig/network-scripts/ifcfg-eth0

  6. Strip out everything but the bare minimum:


  7. Save file, then reboot:


The virtual machine will come back with eth0 raised and a brand new IP address.

If I didn’t know any better I’d say someone is trying to recreate the behaviour that Windows has if you clone it. Each clone of a windows system remembers all the MAC addresses and names of previous NICs but what it means is that when you clone a system and update the drivers then all your names change back to the default plus one. for example if you rename your “Local Area Connection” to “outbound-if” and clone the system when it boots up on a different host with a slightly different firmware revision or different NIC then when the cloned system boots up, “outbound-if” will not be there “Local Area Connection 2″ will be there instead. You can’t change the device name to “outbound-if” because that name is taken by the system. There is no way to change this.

I think the idea is that those settings are tied to the MAC address (hardware). So when the hardware changes you get a default configuration as standard. In my experience I’ve never had to add extra settings and my hardware changes have always been like-for-like. It has always been a nightmare that has taken ages to fix.