Frequently Asked Questions

[Thanks to Bill Studenmund who reported this problem and its solution]

1. If you use HiveBoard with PostgreSQL on Tomcat, then don't forget to copy the PostgreSQL JDBC driver (jar file) onto tomcat directory tree (on Tomcat 5.0.X, it should be copied on common/lib).
2. If you want to run HiveBoard under a non-default SQL user ID, then you MUST log onto PostgreSQL with THAT user ID to create the HiveBoard database and tables.

This message will appear in two occasions:

• You try to connect a HiveBoard client to a HiveBoard server, but the client and server have different versions. In this case, please upgrade to the latest version (i.e. 0.5.0) of both client and server. Please note that server 0.5.0 won't work with any version of client but 0.5.0.
• The HiveBoard server had errors at launch time (yes, the message is a bit misleading here, a better message will be used in a future HiveBoard version). You can try to figure out what happened by taking a look at HiveBoard server log files. The main reason for server failure is improper configuration, the best way to fix it is to check the installation according to the steps described in HiveBoard documentation.

It looks like you have just been bitten by Java JVM bug #4913324!

This problem happens sporadically particularly on multi-core CPU computers (I could never reproduce it on a single core so far).

This problem is easy to recognize: after you have selected an action (from a menu for instance), a message box is displayed, it has a title but no content at all (no icon, no message, no button). Then the HiveBoard client is hanged until you explicitely kill it (through the commands available, depending on your OS).

The bad news is that, so far, there is no real possible fix for this bug.

The good news is that there is one work-around! You just have to open your client.properties file and set the waitcursoreventqueue property to false. This shall remove this bug. The trade-off is that this property will disable the "automatic wait pointer on long operations" feature of HiveBoard . What? You had not even noticed this feature? Then go ahead and use this work-around with no regret!

Note: I could never reproduce this bug under "normal" use of HiveBoard (I had to create ultra-quick automatic tests to find it). I am not 100% sure if this may happen under usual conditions, so if you face this problem, please let me know about it (and give some details about the conditions: which message box was hung, on what OS, what hardware, how many times you got it).

If you want to manage the HiveBoard server by using JMX (with jconsole for example), then you have to:

1. Setup default properties for JMX management in management.properties file that is located in <JRE_HOME>/lib/management directory; this is useful only if you want to use JMX management from a remote computer.
2. Launch the HiveBoard server from the command line: java -Dcom.sun.management.jmxremote=true -jar start.hiveboard.server.jar
3. Launch your favorite JMX console tool (Sun's jconsole will do) and connect to the HiveBoard server. From the console, you can access the "usual" Sun configuration (JVM, memory, threads...) and also jetty MBeans (the embedded HiveBoard server is based upon Jetty Servlet container).

Note that it is not possible to use JMX by double-clicking the start.hiveboard.server.jar file (the com.sun.management.jmxremote system property MUST be set at JVM launch time and the only way to set it is on the command line).