Blogging Roller

Cedric Beust pointed out that the Win Tech Off-Topic mailing list has been discussing recent J2EE vs. .Net benchmarks. I wanted to get the .Net perspective too, so I joined the Win Tech Off-Topic Yahoo Group to take a look. As you would expect, the Microsoft fans have a different view of the benchmarks than us Java bloggers.  I found no great revelations, but reading the thread was somewhat entertaining.  It went something like this:

• Wow, what an ass whoopin
• Yep and one of those $40,000 app servers crashed!
• Rickard says the benchmarks were a scam
• Who the hell is Rickard, why would you trust a dreambean?
• Rickard is an expert, I would trust him more than The Server Side
• Rickard and The Register and anti-microsoft conspiracy nuts
• Lots of talk about the "appeal to authority" argument
• Lots of talk about whether the nickname "M$" is hate speech or not
• The benchmark was flawed, .Net and J2EE performance are similar
• Price/performance favors .Net because the app server is free with the OS
• Look at that anti-Microsoft stuff on Russell Beattie's site, that guy is sick!

Just as rambling and crazy as any thread you'd find on The Server Side. I guess that is why they call it "off-topic." Why do I waste my time like this?  I should have bypassed that and turned my attention first to Ted Neward's nicely balanced article on the topic: The Petstore...Again.

SharpDevelop, an open source (GPL) IDE for C# and VB.Net projects, seems like an inexpensive and fun way to get started with some .Net education. All you need is the .Net SDK and you are off and running, right?

Yep, all you have to do is to download the .Net SDK and start hacking away. But wait, the <a href= "http://msdn.microsoft.com/downloads/default.asp?url=/downloads/sample.asp?url=/msdn-files/027/000/976/msdncompositedoc.xml">.NetFramework SDK is a 137,363,456 byte download. And that is not all, if you want to redistribute your .Net apps you'll also need to download the <a href= "http://msdn.microsoft.com/downloads/default.asp?url=/downloads/sample.asp?url=/MSDN-FILES/027/001/829/msdncompositedoc.xml">.Net Framework Redistributable, another 21,827,584 bytes. And you thought the JDK was bloated!

Andy published a <a href= "http://www.superlinksoftware.com/cgi-bin/jugwiki.pl?AndyReviewsEclipse">review of the Eclipse IDE on the <a href= "http://www.superlinksoftware.com/cgi-bin/jugwiki.pl?HomePage">Triangle JUG's Wiki. I agree with just about everything he says. Like Andy, "I'm not really a 'real' IDE user. I drop into vi and start clacking away like it was second nature at the first sign of the IDE getting in my way. And I use ant to build EVERYTHING." With Netbeans, I was always dropping back into vi. With Eclipse, I never feel that need.

In other Andy news, Andy <a href= "http://www.trilug.org/%7Eacoliver/hackinglog.html">mentioned today on his Hacking Log the possibility of a joint venture with the Triangle Dot-Net Users Group and the Triangle Java Users Group: a Dot-Net vs. J2EE shootout. Now, that sounds like a fun meeting.

Don't let it stop.

<a href= "http://www.raibledesigns.com/page/rd/20021103#the_blogging_roller_has_a">Matt and <a href= "http://greg.klebus.com/blog/index.do?date=20021103#225825">Greg noticed that I've been futzing around with my templates. Yes, it is fun and yes I fixed my background - thanks Matt. I'm trying to understand CSS a little better. I'm still experimenting so expect more changes.

It has been a week now since the TSS published The Petstore Revisited, but I am just now getting around to reviewing the story.  

On Monday Oct. 28, 2002, The Server Side published (TSS) The Petstore Revisited.  This benchmark report compared Sun's example J2EE Petstore application with some performance revisions added by TSS to a Microsoft written dot-Net clone of the Petstore.  Respected open source Java guru Rickard Oberg quickly reviewed the report, found many flaws, and published a rebuttal the very next day. Rickard updated his rebuttal with more information and links to other's views throughout the week (but he missed Rafe's comments). On Wednesday he was Slashdotted.  

As the week progressed, the outcry continued, and the full story was revealed:

• the benchmark report was seriously flawed in many ways,
• the J2EE Petstore performance revisions done by TSS were inept at best,
• Microsoft paid TSS to do the benchmark,
• TSS is essentially a Microsoft business partner,
  
• and TSS worked with Microsoft to coordinate the dot-Net victory media spin.

If you take the time to read Rickard's review and the comments that he links to, it is clear that dot-Net did not defeat J2EE, but Microsoft certainly did lay waste to The Server Side's reputation.  The TSS benchmarks taught us nothing about the relative merits of dot-Net and J2EE, instead we learned about incompetence and betrayal.

If you are concerned about the relative performance of dot-Net and J2EE, read the Pet Shop Boys write-up at Larry O'Brien's Thinking In Dot Net weblog:

Looks like FreeRoller is still up. How about that!

Anthony has found and fixed what looks to be the source of the FreeRoller crashes. I'm keeping my fingers crossed and I will keep you posted.

I've been pretty busy lately, with a software release at work coinciding with a software release at home and the ongoing FreeRoller difficulties. I need to burn some vacation time so I'm going to take a couple of days off and relax around the house. I'm not sure if this means more or less blogging. Probably more.

We've been having a helluva time tracking down the problem that is causing <a href= "http://www.freeroller.net">FreeRoller to crash two or three times each day.  By "crash" I mean that FreeRoller gets slow, stops responding to hits on Roller's Velocity-driven pages, and eventually seems to lock up Tomcat. Once Tomcat is locked up, you will started getting "Document contains no data" errors in your browser.

Here are some of the things I've been working on to resolve this problem:

Ensure that resources are released

Roller uses Castor JDO and the Velocity DataSourceResourceLoader for all database access, so Roller has to trust those components to use and release database connections properly.  Every time Roller gets a database connection (a JDO Database object), it uses that connection within a try block and releases the connection within a finally block.

FreeRoller uses Roller on Tomcat with DBCP connection pooling <a href= "http://jakarta.apache.org/tomcat/tomcat-4.0-doc/jndi-resources-howto.html">via JNDI and the MySQL JDBC drivers. We have not be able to figure out how to configure DBCP to get debug informaiton on the number of database connections in use. If you know how to do this please drop me a line.

Ensure that exceptions are being handled properly

There were a number of places where exceptions were being blown out to Tomcat instead of being handled properly by using a response error code.  There was also one place where an exception was being thrown from within a catch block.  None of these should cause a crash, but I was guessing that perhaps one of them was tickling a Tomcat or Java VM bug.

Look for infinte loops and infinite recursion

I found and fixed an infinite loop in the error.jsp file which used itself as its own error handler.  That bug was very easy to find due to the big stack trace and stack overflow exception in the Tomcat logs.  

There could be other problems with infinite loops and recursion, but I have not found any evidence of these problems in the FreeRoller logs.  Weblog templates and weblog entries are treated as Velocity templates, so it possible that a weblog author could introduce an infinite loop or infinite recursion.  Roller tries to protect against this in the Roller macros.includePage() and macros.showWeblogEntries() directives, but we can't offer protection against a truly malicious user.

Test under heavy load

FreeRoller runs on Tomcat 4.0.6 and MySQL on Linux.  So, I setup Tomcat 4.0.6 and MySQL on my Redhat 7.1 box and tried some stress tests.  I tried using JMeter, but it kept on running out of memory.  I eventually found Openload and set it up to hit http://snoopdave/roller/page/test1 (a pretty heavy page) and http://snoopdave/roller/rss/test1 with 5 clients on each.  I configured Roller with a session and page-cache timeout of 2 minutes so that the database code gets a workout.  I ran for several hours at heavy load with no ill effects and the number of JDO Database objects never exceeded 20.

Even with a new 0.9.6.3-pre1 build of Roller that addresses the issues above, FreeRoller still had a crash today.  I'm baffled.  If you have any suggestions for testing or debugging tools or approaches, please let me know.  Free feel to browse the code ;-)

I updated the <a href= "http://rollerweblogger.org/userguide/roller-ug.html">Roller user guide for 0.9.6. It is not perfect, but it is better than nothing. I also posted the <a href= "http://www.rollerweblogger.org/javadoc/index.html">Javadocs for Roller.

With the arrival of spam in the world of weblogs, can the arrival of Microsoft be far behind? Spam arrived on the internet in 1994 and the very next year in December 1995 Microsoft announced its massive assault on the internet with ActiveX, Java, a web browser, etc. It won't be long before we see <a href= "http://www.dashes.com/magazine/backissues/microsofts_weblog_software.php"> Microsoft Weblogger 1.0 down at Best Buy.

I feel bad about the FreeRoller outage. I still don't know what the problem is, but I can't help but blame myself. I really appreciate the fact that Anthony Eden is hosting Roller at FreeRoller and giving me the opportunity to see how Roller scales with lots of users. I hope Anthony and everybody else will have patience as we try to track down this problem. I miss those blogs!

Apparently some very evil spirits still lurk deep in the heart of Roller. My attempts at exorcism have failed. I soaked error.jsp in holy water and was able to save FreeRoller from the grasp of the stack-overflow demons, but there is still something wicked going on. FreeRoller is still crash thrashing up and down. Maybe a fresh new build (0.9.6.3-pre) will rid us of the goblins.

As Ed Savage wrote on the Triangle JUG mailing list: this is big news. According to the <a href= "http://www.bizjournals.com/triangle/stories/2002/10/28/daily39.html">Business Journal article, Borland is paying $185 million in a combined cash and stock deal.

Borland has been on a bit of a shopping spree lately. They recently bought StarBase which gained them the StarTeam issue tracker/source code control system and the CodeWright editor.

At the Eclipse meeting last night, I noticed that both TogetherSoft and Borland are on the Eclipse steering committee. Will Borland stop TogetherSoft from rehosting inside Eclipse or is Borland considering rehosting all of it's various IDEs and dev tools inside Eclipse? Will Borland help IBM make Eclipse into the universal multi-language tools platform?

I've been meaning to comment on this for a couple of days. Niel Eyde recently posted some words about JSP and MVC. Some people criticize JSP because it allows code-on-page. I understand that concern and I've got scars from dealing with frightening model 1 JSP code before, but I agree with Neil that JSP is not evil.

A frequent criticism that I hear about Java Server Pages (JSP) is that it doesn't strictly separate the presentation logic from the business logic, and there seems to be many solutions developed that aim to force this strict MVC separation. As developers, why should we depend on technology to force adherence to the goals of MVC? It is certainly possible to develop JSP pages that adhere to principles of MVC. If model or controller logic creeps into a JSP page, it our own fault. [Niel Eyde on JSP and MVC]

Greg recently moved off of FreeRoller to his own site at http://greg.klebus.com/blog. He left FreeRoller, but apparently not because he disliked Roller. Thanks Greg!

[thanks to] Dave Johnson and Roller contributors: Matt Raible, Lance Lavandowska. I must admit Roller v. 0.9.6.x seriously kicks ass. It's awesome. Great work. I had been tempted to install Roller on my webserver, but finally decided that the simpler application would be more appropriate.

FreeRoller has been up and down for the past couple of days. After looking at the logs, I that believe Roller has been going into an infinite loop of recursion because Roller's error page error.jsp declares itself to be it's own error page. When an error occurs on error.jsp, error.jsp is called where an error occurs and error.jsp is called where an error occurs and so on and so on. I'm not sure what the first error is or what the error on error.jsp is, but I've given FreeRoller a new error.jsp that should have better behavior. Maybe the new error.jsp will also reveal the root of the problem.

Last time I checked, it appeared that the Triangle Dot-Net Users Group was stillborn. But now, Tri-NUG appears to be cookin'. Recent speakers included Jeff Prosise and Jeffrey Richter, names I remember well from the bad old days of MFC and Windows API.

The RTP Websphere Users Group meeting meeting tonight was great. Eclipse Product Manager John Kellerman gave an excellent talk on Eclipse and covered the Eclipse history, philosophy, capabilities, plugin architecture, and future plans. He went into great detail in his presentation and in his demo before he ran out of time. I've been using Eclipse heavily, both at home and at work, for over a month now and he was still able to point out lots of features that I had not even noticed. Great stuff. Go IBM!