July 2006 Archives
Not so much flying this weekend - the weather was just too darn hot. Speaking of hot, there's been some push to get the internal mail server upgraded at work.
Fearing yet another Cyrus migration, I took at look at Dovecot.
Wow.
I've been a long time Cyrus mail storage fan. It is a very robust and scalable system. Messages are indexed for the main headers at the delivery time, making it possible to open huge monster mailboxes in under a second (with a good imap reader, at least). The actual storage is one message per file (eat those inodes!), with indexes; and the indexes are stored at delivery time.
That said, Cyrus is a pain in the ass to set up. No two ways about it. Even harder is upgrading, particularly any time they change the database format. In the past I've found it faster to dump the databases entirely and let it scan the hard way to rebuild. In general, not a lot of fun.
So, for this server upgrade over the weekend, I took a look at dovecot. From where I stand, there are some benefits for our group:
- You can manage the users with a single text file. Sure, you can do various databases, ldap, write your own adapters whatever.. but.. for us, a single text file is great.
- Since we're configuring it with jst a text file, we can drop building (and later depending) on all those libraries. No gdbm. No Berkeley DB. No SQL server. No LDAP. The only thing we linked in was SSL.
- Storage: more "normal". Choice of maildir or mbox. Hell, use them both.
- Cyrus's speed advantage is from precaching headers. Dovecot _can_ do this, but the production method right now is to cache headers (not precache). In practice, this looks to be just fine. I can still open that huge 50,000 message mailbox in a second.
Another neat thing, is we've got it set up so that we have maildir for folks, as the standard location; but with the right path via imap, someone can just as easily open a legacy mbox format mailbox. I'm personally storing my "backups" mailbox as mbox to limit the impact on the server, since it is not routinely opened.
If you're looking at imap servers, or even pop servers, do yourself a favor and spend a bit of time looking at Dovecot.
urls:
http://www.sportys.com/faatest2/default.cfm
This will let you do a sample run of 60 questions, time you, grade you, and then let you see all the right/wrong questions and answers. Also, all the needed materials are available (you don't need your own private copy of booklet FAA-CT-8080-2E of your very own.. but.. it helps).
http://www.logshare.com/log.jsp?email=jfesler@gigo.com
- 36 flights over 11 unique routes and 4 planes
- 64.9 hours total, 6.1 of those night time, and 20.9 of those solo.
- 183 day landings, 13 night landings.
Time to start studying like a mad man, and pass the written test (scheduled for the 28th)..
.. and leap with your hands out in front of you.
Today was my "cross country" solo. 260 miles round trip, no instructor. Just me, the plane, and every other plane out there. Even the birds stayed out of the way!
Soft fields present an interesting challenge for an airplane, especially a tricycle gear airplane. You have to adjust your takeoffs and landings such as to keep all the pressure you can off the front nose.. so that it doesn't sink into the muck.
Landings, you come down on two wheels, and intentionally keep the front end of the plane from touching down, and stay off the brakes. The field you're landing on will slow you down plenty.
Takeoffs, you try and keep the pressure off the nose gear - and actually raise the nose gear as early as you can. Looks funny as hell. Trying to fly that way? annoying as hell. Once you get enough speed, you get off the ground, but stay in ground effect - getting off all 3 wheels as soon as you can. You can do this at low speeds, but if you try and actually move away from the earth, you'll loose the the ground effect, and come crashing down. So, you have to build up spee d while floating _over_ the ground until you have enough for a proper departure.
I also practiced "short field" landings today, per the book. According to the book, I can land that plane in 550 feet. Currently, I'm landing it in about 750 feet. We'll see how close I can get to what the book says. In any case, I'll learn my personal minimums, on what I can safely land with.
Tomorrow: cross country, Sacrmamento to Chico to Redding to Sacramento, solo... assuming the weather is kind.
During preflight, one of the things you do is drain fuel from all the sump points, and make sure it smells flamable and looks blue. (100 low-lead aviation fuel is dyed blue). Why?
Condensation. Water vapor is in the air around us. The amount of water that can be held depends on the temperature. Once the air chills below that point, the water will condense, turn to fog/cloud, or in some other fashion re-appear into a for we know it as.
Unfortunately, fuel tanks, if not kept topped off, can have a fair amount of air in them. The plane I had today, had 13 out of 40 gallons filled, the rest of the space was air.
I did not find water in the wing drain points (where fuel is stored), but I did get water at the engine fuel drain. Not a huge amount, but some. It looked like a bubble that sinked to the bottom of the fuel sample. I drained and sampled repeatedly until all the water was out, and then I drained a bit more, to be certain.
The plane did run rough for the first moments, but cleared up quickly. More importantly, I didn't feed the water into the system on takeoff.
Otherwise, I might have emulated
this accident in Rancho Murietta. There was a writeup in one of the flight magazines recently, by the instructor who was involved in that accident.
So, for today, I got to see a real life example of what you preflight so dillegently, and, I had ten touch-n-gos that didn't suck. All in all, an interesting but good day.
