Month: February 2005

SQL1224N A database agent could not be started to service a request,
or was terminated as a result of a database system shutdown
or a force command. SQLSTATE=55032

This was the error I kept getting without being able to find out why. A couple of things I have tried:

• Upgrading JBoss from 3.2.6 to 3.2.7 (which made me try point 2 since 3.2.6 doesn’t run under Java SDK 1.3.1 even though it says so on jboss.org)
• Downgrading from Sun Java 1.4.2_07 to IBM Java 1.3.1
• Upgrade to various DB2 Fixpacks culminating with Fixpack 13

I found the solution as technote 1044438 on IBM’s website (I also took a PDF print to make sure I could find it in the future). The technote only mentions AIX but the solution works on Linux as well.

The solution is to create another DB2 node pointing to the same DB2 server and cataloging the database using this “dummy” node instead of directly using the local protocol. I have not idea why this solves the problem.

Make sure the STLinks files can be correctly loaded in iNotes. This depends on setting the correct hostname in the “Fully qualifies hostname” field on the server document of the Sametime server and making sure the protocol used to fetch files from the Sametime server using HTTP is the same as the protocol used for iNotes. This means that if iNotes runs over HTTPS, iNotes will also try to fetch Sametime files from the Sametime server using HTTPS. If the Sametime server doesn’t support HTTPS set the iNotes_WA_SametimeProtocol notes.ini variable to http: (the trailing colon is important). This will tell iNotes to use HTTP instead of HTTPS in the above example.

We use a lot of XML fetching in our web applications to fetch information from the server and processing it with JavaScript afterwards. What we normally do is to have a target division hardcoded in the correct location on the page and use JavaScript to add content.

The below example is taken out of context and it utilizes a custom JavaScript object that does all the dirty work and calls the processXML(boolean success) method once the XML is ready for processing. The interesting part is where the HTML is dynamically built using the createElement(String name) and appendChild(Element e) functions.

Searcher.prototype.processXML = function(success) {
  // how did we fare ?
  if (!success) {
    window.status = "Something went wrong while getting the XML while searching - aborting...";
    return;
  }

  try {
    // loop databases
    var databases = this.getElementsByTagName("database");
    if (null == databases) return;
    for (var i=0; i<databases.length; i++) {
      // get next database
      var database = databases[i];

      // get the title and path of the database
      var db_title = this.getTagValue(database.childNodes, "title");

      // get all results for database
      var results = null;
      for (var j=0; j<database.childNodes.length; j++) {
        if (database.childNodes[j].tagName == "results") {
          results = database.childNodes[j];
        }
      }

      // get result count
      var r_count = results.getAttribute("count");

      // create element for db title
      var elem_db = document.createElement("div");
      elem_db.id = "database" + i;
      elem_db.innerHTML = db_title + " (" + r_count + " resulter)";
      elem_db.className = "searchresults_database";
      elem_db.onmouseover = changeCursor;
      elem_db.onmouseout = changeCursor;
      elem_db.onclick = clickDatabase;
      this.pResults.appendChild(elem_db);

      // create element for result section (collapsed)
      var elem_results = document.createElement("div");
      elem_results.id = "results" + i;
      elem_results.className = "searchresults";
      elem_results.style.visibility = "hidden";
      elem_results.style.display = "none";
      this.pResults.appendChild(elem_results);

      // loop results
      for (var j=0; j<results.childNodes.length; j++) {
        if (results.childNodes[j].tagName == "result") {
          // get next resultsitem
          var r = results.childNodes[j];

          // get title and key
          var r_title = this.getTagValue(r.childNodes, "title");
          var r_key = this.getTagValue(r.childNodes, "key");
          var r_link = this.getTagValue(r.childNodes, "link");

          // add link anchor
          var elem_link = document.createElement("a");
          elem_link.href = r_link;
          elem_link.className = "searchresult";
          elem_link.appendChild(document.createTextNode(r_title));
          elem_results.appendChild(elem_link);
          elem_results.appendChild(document.createElement("br"));
        }
      }
    }

  } catch (exception) {
    // some error occured
    window.status = "Exception caught in Searcher.processXML(): " + exception.message;
  }
}

One thing I found out is that the Microsoft XML component allows for asynchroneous fetching. This allows you to show a progress bar or similar while loading and processing in the background.

function go() {
   // set to fetch asynchroneous
   this.pXML_data.async = true;
   // set callback method
   this.pXML_data.onreadystatechange = waitForXml;
   // go fish !!
   var success = this.pXML_data.load(url);
}

function waitForXml() {
   // wait for document to be completely loaded
   if (gXML_handler.getXmlIsland().readyState == "interactive") {
      // document has been loaded - process it
      processXML();
   }
}

function processXML() {
   // do processing...
}

Just a reminder to look into mod_rewrite:

• mod_rewrite guide @ Apache.org
• mod_rewrite manual @ Apache.org
• mod_rewrite info @ Apache.org
• mod_rewrite info @ pbblog
• mod_rewrite info @ pete frietag
• mod_rewrite info @ Migs Paraz
• mod_rewrite info @ Moertel

When configuring Apache 2.0 before compiling (make) and you are enabling mod_ssl you sometimes get an error saying that the OpenSSL headers cannot be found. To fix this make sure to configure using the –with-ssl=<path to OpenSSL installation> parameter.

Trying to configure JDBC for DB2 v.8.x I had an error messages that just wouldn’t go away:

java.lang.UnsatisfiedLinkError: no db2jdbc in java.library.path

The problem appeared to be configuration issues and the fact that DB2 v.8 ships with a new universal JDBC driver. There are a number of things that are different about the new driver:

• New driver jar-file (db2jcc.jar)
• Requirement for a license jar-file (in my case I guessed that the required license jar-file was db2jcc_license_cu.jar since it was the only one in /opt/IBM/db2/V8.x/java)
• Change driver class name to

  com.ibm.db2.jcc.DB2Driver

• Make sure the native library are in the library path if using the type 2 driver (JDBC against a local DB2 installation):

  export DB2PATH=/opt/IBM/db2/V8.1
  export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$DB2PATH/lib
  

The JDBC URL syntax is the same as previously.

If you forget adding the DB2 directory to the LD_LIBRARY_PATH you’ll get an error message saying:

com.ibm.db2.jcc.b.SqlException: Failure in loading T2 native library db2jcct2

To install DB2 manually follow the below steps:

1. Download DB2 v.8.x for Linux (I downloaded v. 8.1 Express) and transfer the gzip’ed tar-file to the Linux box.
2. Unpack the distribution to the /tmp directory:
   

   tar zxf DB2ExE82_Linux.tar.gz -C /tmp
   

3. Change to root and go to the unpacked DB2 directory and start the installation:
   

   cd /tmp/Linux
   ./db2_install
   

4. Select the package to install – I chose “DB2.EXP” since I was installing DB2 Express.
5. Create an user for the DB2 instance and an user for the administration server (make sure the users belong to a valid group and that the users have a home directory):
   

   useradd db2inst1
   useradd db2as
   groupadd db2inst1
   groupadd db2as
   mkdir /home/db2inst1
   mkdir /home/db2as
   chown db2inst1:db2inst1 -R /home/db2inst1
   chown db2as:db2as -R /home/db2as
   <associate the group with the user e.g. using yast under SuSE or manually editing
   the /etc/groups file if you are so inclined>
   

6. Create an instance:
   

   cd /opt/IBM/V8.1/instance
   ./db2icrt -u db2inst1 db2inst1
   

7. Create the administration server:
   

   cd /opt/IBM/V8.1/instance
   ./dascrt -u db2as
   

8. Set the environment for the db2inst1 user by adding the following to the users ~/.profile:
   

   ./sqllib/db2profile
   export DB2INSTANCE=db2inst1
   export DB2PATH=~/sqllib
   export PATH=$PATH:$DB2PATH/bin
   

9. Log in as the db2inst1 user and check the installation by creating the SAMPLE database:
   

   su - db2inst1
   db2sampl
   

Sources:

• Google
• informit.com

Got to look into configuring mod_gzip for our web severs.

Info on configuring mod_gzip

The problem was that we got confused with the difference types of server names needed in different places. If you got confused what are you doing setting up Sametime servers one might ask, but in our defense this is the first time we experienced this problem. Whether we have just been “lucky” I don’t know. Anyways…

When setting up Sametime for an user you need to input the Sametime server on the person document in the Domino Directory and the Sametime server on the users location document. The problem turned out to be that the server name on the person document should be the fully qualified name of the server (e.g. Sametime1/XYZ/Example) and the server name on the location document should be a resolvable hostname or IP address to the server (e.g. sametime1.example.com / 10.0.0.1). Once we found this out configuring it was a breeze using the Domino Administration and policies.

So much for easy to read documentation… 🙂

Update (18 February 2005): Apparently not the only one being confused – a technote has been published.