Thursday 5 January, 2006

Ok, so I haven't gone to sleep yet. Still up working this AOP thing. Once i got it compiled i wanted to put it to use. Got some of the key code in place, although hastily. Now I need to tweak and test on some of the machines around here.

Update 2:03AM: Still up tweaking. Gotta go to bed.

Update 8:12AM: Up with the Avers.

Update 8:58AM: At the hoffice in the 180 Room, ready to start testing clustering.

Update 11:55AM: Clustering didn't work :(

Heard from Dana via email. Touched base with John. Working with Kendra on a client private label. Wrangling html/css.

Getting into that unfocussed part of the day where things hit me from all angles. Got to slow down and take one at a time.

Update 2:30PM: Now I'm hungry. Taken 476 calories on board thus far. This is more than yesterday at this time. But now I'm hungry. I've been filling between meals with apples and bananas.

Just remembered that I have a ton of ava photos to upload from the holiday season. Gotta get to that soon.

Clustering still not working. But Kendy and I had a good IM session getting the client's PL almost set up. I'm turning my attention back to clustering now.

Update 3:22PM: Oooh, this is some really cool Xxxt. I was debugging the lack of clustering activity... wondering whether it was a code bug or network connectivity issue. I fussed with the router a bit, setting up a static route for multicast traffic (unnecessary). I then found a test utility in the jgroups library I'm using. It's command line. You launch it as a listener on one machine and then as a sender on another. When you type in one you see it in another. And it worked! So, geek points to me. Nothing much more than text chat, but still. Then I hooked the listener up on the clustering port I had chosen and, as if by magic, I started to see "RegerCom-TreeCache-Cluster" messages flying back and forth! WahoO! So I have two instances of the Reger.com app on different machines talking to one another behind the scenes using multicast.

So networking issues are out. Which lets me move on to a new level of debugging. The clustering thing still isn't working. But it's been very cool seeing the messages going back and forth.

Ok, back to debugging.

Update 3:28PM: Even cooler: run the listener on both sides of the wire and watch the fireworks!

Update 3:30PM: Re-reading the above updates: maybe this explains why I didn't get laid in high school.

Update 3:35PM: That was fun. Now I'm stumped again. Dammit.

Update 5:06PM: Actually R'dTFM. Helped. Saw a huge hawk... five or six foot wingspan outside. Nice. Sun is down so I can see my screens better now. Sunny days aren't optimal for computering in the 180 Room.

Update 5:25PM: Heather on her way home so I only have a short period left to work. Made some progress... the AOP engine is definitely kicking in now... it wasn't before. But now I'm seeing some real bugs and issues. Trying to work through them. Nothing is ever plug and play.

Update 8:42PM: Went for a run. Ate some vegetable and beef stir fry. Watched some tv with Heather and Ava. The 180 Room is now cold and it's rather windy out.

Update 8:58PM: Just as I was starting to fall into the dark hole of despair... a little hope. Broke through the bug and am now moving forward again. Roller coaster, this coding thing.

Update 11:51PM: Aw, it's talking but not sharing the sessions. Frustrating. I need to bump the debug level higher to see what's happening.

Pawan's got Java 1.5 installed and running.

I'm tired and I have to get up early (for me) for an 8AM haircut.

Update 12:06AM: Well boy wow howdy! I set log4j debugging to DEBUG on the org.jboss.cache pakage and a single request generated 23.6Mb of debug file... text. Now that's a lot of detail. It could keep me busy until 3AM. I'll probably just peruse it a bit and then go to sleep.

Update 12:40AM: So... damn... close. And yet so far.

Update 12:44AM: Windy and cold out. Got a small ceramic heater at my feet keeping me warm in the 180 Room. Ok, I give. I'm beat. I'm going to bed.