One of the things I’ve realized with this new position is that I am my own worst taskmaster, driving myself to work longer hours in tightly focused stretches of time rarely punctuated by breaks. I suppose on some level that I feel like I need to be even more productive because of the absence of “face-time”, that there is no boss leaning over me making sure that I at least have the appearance of being busy. In contrast, though, I really am enjoying the work and the challenge that it presents, so I often feel that itch in the back of my brain to tray and solve the puzzle before I go to bed.
What have I been working on? Well, half the week was spent training Monit to play nice with Mongrel and I am decently confident that it works as advertised in the test environment. This afternoon we did a test deploy with Capistrano nesting it between Monit stop and start statements and everything appeared to work without a hitch. The challenge we faced with Monit in our environment was that we are unable to actually issue Mongrel starts and stops inside the config file. The solution was to take those statements and drop them into bash scripts, so at the moment I have an kludgey but operational method of fourteen scripts for seven mongrels (one start and one stop). When I get a moment, I plan on cleaning them up and making a single one that executes with variables, ie $ monit-mongrel stop 8001 but at the moment I am epically lazy. If I have the time I would like to figure out what exactly it is about the environment that doesn’t like mongrels being started or stopped inside Monit.
Half of yesterday and all of today I have been pounding my head against a nail studded board trying to get secure replication rolling inside EC2 for our MySQL boxen. The masters and slaves (yes, the developers on the crew with a more PC sensibility have chided me saying that the correct terms are primary and secondary. Fine we can meet in the middle with boss and underling) fire up fine and do what they are supposed to except actually perform replication of any shape, form, or fashion. To pipe them together I went the Stunnel route–could not for the life of me get SSL in MySQL to actually do anything–and I know that something is happening because the moment I issue a SLAVE START; command this shows up in the stunnel logs on the master: 2007.08.03 15:46:10 LOG5[13077:3083316112]: localhost.3306 connected from xx.xx.xx.xx:36769. I’m thinking that possibly it is how I set up the replication account permissions on the master, GRANT REPLICATION SLAVE ON *.* TO ‘replicantsarepeopletoo’@'%.mydomain.com’ IDENTIFIED BY ‘s3KrEtpa5Sw0rd’;. Taking shot in the dark, since I am tunneling the traffic it likely should just be ‘replicantsarepeopletoo’@'localhost’ so when I’m feeling a little less punchy I’m going to take a look at that again but after twelve hours I pretty much hate MySQL and EC2 at the moment.
What I haven’t been doing is taking pictures, writing, reading something other than man pages and long-winded newsgroup threads, and really listening to some of the new albums I just picked up (Nicole Willis & The Soul Investigators – Keep Reachin’ Up and Red Bumb Ball: Rare and Unreleased Rocksteady (1966-1968) are fucking amazing albums though). Hopefully, I’ll find my stride soon and build a sort of groove where I’m not pushing myself so hard that I’m dreaming about how the company abandons it current market focus and I’m forced to look into re-architecting the mongrel cluster for their plans to launch a fried chicken franchise. Yeah, I do need more sleep.
Tags: EC2, Mogrel, Monit, MySQL, Reflecting, Stunnel, Thinking, Work
Fried chicken is good!
Fried chicken is good but how is it at scaling?