MySQL Cluster 5.1

I've been researching MySQL cluster today and my brain is on a bit of information overload so I thought it would be wise to jot down some key points I have learned thus far.

MySQL cluster differs from a traditional MySQL installment because it is designed to work across multiple machines. In the traditional setup you install MySQL on one node and have several slaves that replicate the data and serve as backup servers in case of failure. In the cluster scenario there are a minimal of 4 machines (technically 3, but if you care at all about your data you will need 4). Machines fall into one of three categories:

1. Management Nodes
Management nodes are important for defining the roles of each of the other nodes in your cluster. Here you setup config files that the other servers read to understand how the cluster is setup. You would need at most two of these, one just sitting there in case the primary fails. If the primary does fail, it will not cause the cluster to stop working.

2. MySQL Client Nodes
Client nodes are the machines that actually perform all of the queries received from clients. You can setup as many of these as you load demands, preferably at least two in case of failure. In MySQL 5.1 these nodes do perform caching, but there are some gotchas that you should read in the documentation.

3. Data Nodes
The nodes use the NDB storage engine to replicate your data across multiple machines. This is layer I have the most questions about and I'm trying to find out more. Prior to 5.1 all database data, including indexes was stored in memory. So if you have large tables you need to make sure your data (and indexes) will fit into the available memory. A new feature of 5.1 is the ability to store only the primary key information in memory and store the rest of the data on disk. This will help a lot of really large data sets.

So if you had 2 management nodes, 2 client nodes, and 2 data nodes any of the nodes in the cluster could be shut down and there would be no loss of uptime or data.

A nice feature added to 5.1 is the ability to replicate clusters. So if you are paranoid and afraid that an entire cluster could go down then you could fail over to the secondary cluster. I didn't read the entire doc on this topic, but it seems complicated and probably overkill for most situations.

It's important to understand how indexes are implemented in the cluster scenario. Primary keys are implemented as a hash index type. They do this so they can partition your data based on the hash. This brings up two key points that must be understood when looking into the clustering:

The following queries will work great and be very fast (from the performance pdf available online):

SELECT * FROM tbl WHERE pk = 5;
SELECT * FROM tbl WHERE pk = 'ABC' or pk = 'XYZ';

The following queries will not use the primary key index:

SELECT * FROM tbl WHERE pk < 100;
SELECT * FROM tbl WHERE pk_part1 = 'ABC';

The last example is especially important to understand. If you have a primary key that uses multiple columns (pk_part1, pk_part2) you must specify the value of both columns in order for the index to be used.

Ordered indexes are implemented using T-tree's instead of B-tree's. This is for memory saving purposes and the need to not have to worry about disk-seek times (everything is in memory). This really should make any difference to the client but I mention here because I thought it was a great idea.

Data retrieval in a clustered scenario is very important to understand. If you do a query based on the primary key, the client node can go directly to the data node that has your row and retrieve it. If you do a query which results in the client having to use an ordered index, the client must query in parallel every data node in the cluster. This may be really fast, but would result in much more network traffic than a primary key lookup. The worst kind of query would be one where the client node must do a full table scan. In this scenario the query must be sent to all nodes. Prior to version 5.0 the "where" clause was not sent to the data nodes so that meant that the entire data set was returned to the client for scanning. If you had 2 million rows, all 2 million would be sent over the network for scanning. Since 5.0 you can enable an option to perform the "where" on each data node and only return the rows that match the where clause.

I think picking a primary key is especially important when implementing a table that will reside in the cluster. Pick the wrong key and I think your queries will be much slower than the traditional database. On that note, it's not uncommon for a clustered database to be slower than a traditional database. If you look at everything that must happen to perform a query in the clustered scenario that makes sense. But I think the clustered database kicks but in the ability to maximize throughput.

As I mentioned earlier, my main questions so far are in regards to the data nodes. I haven't found any documentation that covers how data is spread across the nodes. For instance, if I had a configuration with 50 data nodes, and I have a database with 100 million rows, how much of the data would be on each node? Is the data triple replicated? Double replicated? Which data nodes are safe to shut down at the same time? Does anyone out there know or could point me in the right direction?

Cool GUI Tools for MySQL 5.0

I just found these tools for 5.0, awesome stuff, but I'll probably still do most tasks the old fashion way. :)

http://dev.mysql.com/downloads/gui-tools/5.0.html

MySQL Partitioning

Bill asked me to look into the new paritioning features with MySQL 5.1. I have to say this has to be one of the best features I've seen added to MySQL in a long time. Check it out:

http://dev.mysql.com/tech-resources/articles/mysql_5.1_partitions.html


Finally the admin has some kick-ass control over how the data is split across files and disks. And it makes pruning old data super-easy.

College Football is broken, here is how I would fix it

I love college football. It's infinitely more interesting than pro football (unless you play fantasy football, which I don't). But it could be even better, here is what I see as college football's biggest problems:

1. Rankings
2. Every game matters too much
3. Out of conference scheduling
4. No playoffs

Rankings matter too much, we shouldn't care about Coaches Polls, AP Polls, computer polls, etc. There are too many polls and none of them should matter when deciding a national champion.

It's really tough for a one loss team to win a national championship, which means colleges that are smart will schedule as many "Northeasterns" as possible. Ideally, colleges should want a tough out of conference schedule to get them ready for the post season.

Without a playoff system, the teams that are extremely lucky and highly thought of in the polls play in the big game. Nothing is decided on the field and there will always be the one or two teams that make tremendous strides during the regular season but have no chance of proving they are the top team in the country.

Here is how we fix this mess:

1. The winner of each major conference (there are 7) gets an automatic bid into the playoffs. This does several things right out of the gate: we no longer care about polls, we can schedule tough out of conference games, and losing one or two games doesn't really matter.

2. The press still gets to pick the 8th and final playoff spot from the remaining conferences and independent teams. (So yes Notre Dame fans, it's still fixed for you.)

3. Force each team to only play 10 games (plus 1 for a conference championship if needed). This will mean that the football season won't last any longer than it currently does (don't forget that they currently take over a month off before the bowl game). And only 2 teams will play an additional 3 games max, so we are talking 14 games maximum, and some teams currently play 13.

4. We can still have several bowls to satisfy the other teams that didn't make the playoffs but had good years. I think the fans would still go to these games and they would still generate lots of money. But let's get real and reduce the number of bowls in half.

Amazingly enough, college basketball has everything right. Polls don't matter that much, they have a great playoff system, and teams schedule tough opponents to HELP their chances of making the playoffs. What an amazing concept!

MySQL: InnoDB vs MyISAM tables

MySQL gives the user several different storage engine options when creating a new table in a database. The two primary options for traditional databases are InnoDB and MyISAM. I won't go into the history of each engine, but I'd like to discuss when you would choose one versus the other. InnoDB has several features that MyISAM does not provide:

http://dev.mysql.com/doc/refman/5.0/en/innodb-overview.html

So why would anyone ever use MyISAM? First, it's been around longer and is very reliable. People are also more comfortable with the one file per table concept. And maybe most importantly it is the default engine when you create a new table. I've used both and like both, and when I was testing MySQL 5.0 over the last few days I decided to put both to several speed tests on a rather small table (45,000 rows). Each row contained about 470 bytes of data. Every test I ran showed that each database had similar performance. I was hoping that one would be superior to other but it just wasn't the case. On the other hand, my table size was pretty small. And I did discover that with large text based primary keys, MyISAM was much faster. But when the key size was reasonably small InnoDB had slightly better performance. I'm going to try to insert a tera-byte of data into each database and see which one performs better. I love all of the different options available for InnoDB storage that are available so it should be fun to try out different parameters and see which ones perform best.

National Championship Predictions

OK, after one weekend in college football I've seen enough to know who has a legitimate shot at playing for the national championship.

1. Ohio State

A great team with two Heisman candidates. A tough opening schedule, but if they survive at Texas and can hold off Penn State in week 4 there will be no stopping this team.

2. USC

All of their "tough" games are at home (Nebraska, Arizona St, Oregon, California, Notre Dame.) Plus they don't play a conference championship game.

3. West Virginia

Their schedule is laughable and no conference championship game.

4. Miami

Underrated and an easy schedule.

5. Penn State

A dark horse candidate, but if they manage to somehow beat both Notre Dame and Ohio State they deserve to play in the big game.

Teams that will NOT be contending, but are good teams:

1. Texas

Their schedule is tough, Ohio State, Oklahoma, and Nebraska. If they do somehow manage to beat all of those teams they deserve a shot in the big game.

2. Notre Dame

I can't see anyone beating USC, especially @ USC.

3. Any SEC team.

SEc schedule too tough, plus they play a championship.

4. Florida State.

Overrated.

Football Season Already?

Awesome.

Tailgating starts tomorrow, I’m REALLY looking forward to this, I’ve missed tailgating. But enough about that, here is my prediction for the 2006 Hokie football season, go ahead and call Vegas:

Northeastern – Win
North Carolina – Win
Duke – Win
Cincinnati – Win
Georgia Tech – Win
Boston College – Win
Southern Miss – Win
Clemson – Lose
Miami – Lose
Kent State – Win
Wake Forest – Win
Virginia – Win

Gator Bowl here we come!

I spent zero time on research and analysis but without an experienced quarterback our defense and running game won’t be able to overcome Miami nor will it keep pace with a good Clemson team.

The sad part about the VT schedule is the mediocre opponents out of conference. This schedule is garbage and the AD should be ashamed of himself. The good part about a weak schedule is that everyone stays in good spirits after the game and makes the post-game tailgate better than the pre-game tailgate. I can’t wait!!!

Our first softball game

Tuesday night we played our first co-ed softball game. Our team consists of Webmail.us employees and our friends. It was Bill’s idea to start the team and we wanted to play so that we would have something to do together and have some fun. Of the 19 people on our roster, I’m pretty sure only 5 or 6 have ever played in a real softball game before Tuesday night. To say we are inexperienced would be a gross understatement. However, our team is a very competitive and eager to learn group of people. I have volunteered to coach the team to the best of my abilities.

We have been practicing for several weeks now and the improvement from week 1 to now has been dramatic. Our team is much better at all aspects of the game and they understand terms such as “cut-off”, “backup”, “calling for the ball”, etc.

So game time rolls around on a dreary and rainy Tuesday night at Tom’s Creek Park. As soon as the game before us finishes they are calling for lineups for our game. I decided to keep myself out of the lineup for several reasons, but mainly I wanted everyone to get a lot of game time experience. I hand our lineup in and we proceed to warming up. I’ve never really seen such a short warm up period but in what had to be under 5 minutes the game was starting. We were the away team so we batted first. We quickly went down 1-2-3 and were in the field for the first time.

For anyone who hasn’t been to a softball game lately the pace of the game is much, much quicker than the baseball you see on TV. Almost every pitch is hit and the bases are very short so doubles are very common. There also isn’t a lot of down time between pitches. So at the start the bottom of the 1^st inning their leadoff batter grounds out to second for an easy out. Cool, 1 down 2 to go. Unfortunately the next five hitters get hits and before you can spin around twice in your chair we are down five runs. We finally get the last out right before they bat around.

As the team comes off the field I can see the shell shock on everybody’s face. So I tell everyone to huddle up and listen. I pause for a second or two and everything is real quiet. What do you say in this situation? I decide to crack a joke and say “Welcome to Softball.” Everybody laughs and I think we all relaxed at that point and the rest of the game went fairly well. Like I told my team, if you take out the first inning and the 4 run inning they had because of a two out error we would have won 5-4. :)

As far as coaching goes I think I did an OK job. There will always be things I can improve on, but overall I really enjoyed it. My favorite parts were watching people backup each other, just like I told them to do time after time in practice. Watching people line up cutoffs and actually hitting those cutoffs, just like I taught them to do. And hitting fly ball after fly ball, just like I told them NOT to do. (We’ll keep working on it.)

New (to me) UNIX Commands

I often have the need to compare two files on different servers, now I can do so without having to scp the file and then run diff:

vimdiff scp://webmail1//home/kminnick/tmp.txt scp://webmail2//home/kminnick/tmp.txt

I also often have the need to take a file and split it into several small pieces:

man split

Query Optimization

One of the more enjoyable parts of my job is finding and eliminating bottlenecks in our system. Yesterday I was faced with a problem of a moderate load average problem on our RSS servers. I could tell that our system has been growing and that there must be a query that just wasn't performing as optimally as when we first launched. I know the RSS code fairly well, so I dove in and starting looking for queries that could possibly be the culprit. It didn't take long to find a query that when executed on the live system took over 5 seconds to finish (table names and columns changed):

SELECT s.id, s.u, s.d
FROM s, i
LEFT JOIN p ON s.u = p.u
AND s.d = p.d
AND p.pr = 'RSS_E'
WHERE (p.val IS NULL OR p.val != '0')
AND s.fId = i.fId
AND s.lDItem < i.rTime LIMIT 1;

OK, so I looked at the columns of the tables and two columns (lDItem and rTime) were not indexed. So I added indexes to both columns and ran the query again. This time it actually took 10 seconds to execute! As my 4 year old daughter would say, "What in the hecka in the world?!?" I've never really seen MySQL slow down after adding indexes, but I know the query optimizer logic is very complex, so I decided to help it out a bit by giving it a hint as to which index to use:

SELECT s.id, s.u, s.d
FROM s, i use index (rTime)
LEFT JOIN p ON s.u = p.u
AND s.d = p.d
AND p.pr = 'RSS_E'
WHERE (p.val IS NULL OR p.val != '0')
AND s.fId = i.fId
AND s.lDItem < i.rTime LIMIT 1;

Now the query only takes .48 seconds to execute! Excellent. I also dropped the lDItem index and the query speed stayed the same, so I only really needed the one new index and a simple little addition to the query to achieve a 10x improvement in the query. This little fixed has eliminated all of the CPU bottlenecks on the RSS servers today. This stuff makes me so happy. :)

The point is that query optimization is not always a matter of just adding an index (many times it is, but not always). I had to experiment several times before I got the results that I expected.

MySQL Chain Replication

This took me awhile to figure out, but here is the option you need to configure chain replication w/ MySQL. This is useful for geographically dispersed MySQL servers. From the MySQL documentation:

--log-slave-updates

Normally, a slave does not log to its own binary log any updates that are received from a master server. This option tells the slave to log the updates performed by its SQL thread to its own binary log. For this option to have any effect, the slave must also be started with the --log-bin option to enable binary logging. --log-slave-updates is used when you want to chain replication servers. For example, you might want to set up replication servers using this arrangement:

A -> B -> C

Here, A serves as the master for the slave B, and B serves as the master for the slave C. For this to work, B must be both a master and a slave. You must start both A and B with --log-bin to enable binary logging, and B with the --log-slave-updates option so that updates received from A are logged by B to its binary log.

Little League Dilemma

I've coached little league ( a long time ago), it was fun, competitive, and more than anything my kids all had fun win or lose (but we never lost). Which brings us to a story from Salt Lake that just blew my mind. Read the full details here, but here's a quick synopsis:

Salt Lake, little league championship. Last inning. The best hitter on the opposing team is up with two outs. The kid has already hit a homer and a triple. Any rational coach would think about intentionally walking the best hitter so that he doesn't hit another homer, makes perfect sense to me. One problem though, the kid that hits after him is a cancer survivor. So they walk the best hitter, and the cancer kids starts crying after he gets two strikes. Oh boy. What were they thinking? The kid strikes out, crying his eyes out. The opposing coach is furious.

So, if you were the coach would you have walked the best hitter to get to the kid w/ cancer? I'm not sure I'd be able to do much celebrating if I did that. Maybe the bigger question is what in the world is the cancer kid doing batting behind the best hitter. You don't put Mario Mendoza behind Babe Ruth...Idiot.

How many of you would play for the win in that situation?

Amazon S3 Part 2

OK, so I got a few complaints (you know who you are) about part 1 of my Amazon S3 post being too long and boring. So....I'll keep part 2 short and sweet. Basically, I wrote a backup / restore script for my linux files using PHP and REST (instead of SOAP). The bottom line is that I love the Amazon REST interface. It's simple and easy to use, forget SOAP, use REST.

The only other issue w/ S3 that I've noticed is that it will fail 1 out of every 2000 or so requests. Based on the forums, you simply have to code around this issue. It's easy to do, but I'm not sure why it fails so frequently. But I've never seen it fail twice in a row, so simply re-trying the request after a failure seems to always work.

Blacksox Turning It Around

My baseball team won this past Sunday 20-2 over a pretty good team. I was amazed. Overall we are only 4-8, in 5th place (8 teams total). But, we have improved dramatically over the last 3 weeks. I'm really looking forward to the remaining games. I'm hitting .310, which isn't bad, but it only is 6th best on the team. My brother is hitting .462 with 17 RBI.

I haven't picked up a golf club in a couple of months, I'm just too busy, so I'll probably wait until the fall to pick it up again.

Our company softball team has formed, I've had fun at the two practices I've been able to attend. It should be an interesting first year, I believe our games should be starting sometime soon.

Amazon S3 Part 1

Two days ago I began looking into Amazon's S3 web service. For those of you who don't know, S3 is described as follows by Amazon:

"Amazon S3 is storage for the Internet. It is designed to make web-scale computing easier for developers."

Hmm...web-scale computing made easy...as a geek there's no way I'm not checking this out, so I went ahead and signed up for an account. The signup process was really easy, I just had to break out my Discover card and they promised to bill me monthly, excellent deal. Once I was signed in I had access to all of the documentation and example code. The cost by the way are as follows:

Pricing
* Pay only for what you use. There is no minimum fee, and no start-up cost.
* $0.15 per GB-Month of storage used.
* $0.20 per GB of data transferred.

I figure I'll owe them a few pennies each month, no big deal. I really just want to play around and figure out what kind of cool stuff I could build if I had the time.

I downloaded the .NET C# SOAP example. Installed it, put in my super secret key, recompiled and kaboom, I was up and running. The example showed me how to program the core concepts with S3 via a command prompt. Here are the technical concepts a developer would need to understand:

1. Buckets

A bucket is simply a container for objects. Each user can have up to 100 buckets. This may sound low, but in reality you only need 1 bucket.

2. Objects

Objects are like files, but they have meta-data around them. Meta-data is data about the objects, key/value pairs. You also have to setup ACL's (Access Control Lists) for each object. You can have an unlimited number of objects. At first glance you may be wondering how to organize all of the objects in a bucket. What is really cool is that you can can use any type of delimeter you want to group objects. UNIX people are use to the "/" seperator, Windows uses the "\" seperator, but you can use whatever fits your application.

3. Keys

Every object has a unique key.

4. Operations

Example operations:

a. Create a bucket
b. Write an object
c. Read an object
d. Delete an object
e. List Keys

That's it. Pretty easy concepts to understand, but it's pretty powerful. So the example project showed me the basic concepts but I wanted to build something useful. So I decided to improve "My Internet Based File System" by creating a program that will allow me to

1. Upload a folder to S3
2. View a list of objects in my S3 account.
3. Download the object to my local hard disk.
4. Delete objects from S3.

After about an hour I had a working version. The hardest part was fixing their example code to handle binary files as well as text files. Once I got that it was just a matter of hammering out the code.

Here is a screenshot of the working version (and yes I do design work on the side):


The main problem with this code is that it reads the entire file into memory before sending the object to S3. Not a problem with small files, but if you ever wanted to upload a big mp3 or something it wouldn't work. In order to get this to work, you really need to "stream" the object to S3. But, doing this via SOAP is rather hard. The basic problem is that SOAP is primarily XML going back and forth. You would need to either dig deep and format the objects yourself (not recommended) or use a concept called "DIME Attachments". Good luck finding example code. And a bigger problem for me was that Microsoft switch their "Web Services Enhancements (WSE)" to use MTOM instead of DIME between versions 2.0 and 3.0. I don't really have the time to try to get this mess working. But I have some other ideas I'm going to play with first to see if I can get this working in a simpler manner, stay tuned.

If you are wondering how S3 works on the backend, the API docs give some clues, here is one:

"If the object already exists in the bucket, the new object overwrites the existing object. S3 orders all of the requests that it receives. It is possible that if you send two requests nearly simultaneously, we will receive them in a different order than they were sent. The last request received is the one which is stored in S3. Note that this means if multiple parties are simultaneously writing to the same object, they may all get a successful response even though only one of them wins in the end. This is because S3 is a distributed system and it may take a few seconds for one part of the system to realize that another part has received an object update. In this release of Amazon S3, there is no ability to lock an object for writing -- such functionality, if required, should be provided at the application layer."

Some helpful hints when using mysqldump

I always have to go to the mysql reference manual for this, so I'm going to blog about it so that I'm fewer clicks away and I don't have to re-read the mysql manual every 3 days.

First, mysqldump is useful for a variety of reasons, primarily backup purposes, but it can also be very useful for dumping a very small subset of data. We do this commonly to dump live data to test databases.

Here is an example:

Say you have a database called 'webmail' with a table called 'address' and you want to dump all of the address data for a particular user ('owner') from the live system to the test system to debug a problem. You would run mysqldump in the following manner:

mysqldump --skip-opt --quick --extended-insert --no-create-info --user=testing -p --where="owner = 'kevin@domain.com'" webmail address > tmp.txt

mysqldump by default uses the parameter "--opt" which means:
--add-drop-table --add-locks --create-options --disable-keys --extended-insert --lock-tables --quick --set-charset

Most likely you don't want most of those in this scenario, so that's what all of the options are about.

And to import the data run:

mysql --user=testuser -p webmail < tmp.txt

The most important tip is to make sure mysqldump doesn't add the "DROP TABLE" commands!

Perl

Sometimes perl is a little too forgiving, this little bug has existed for over two years in one of my programs and I just now spotted it...

if(length($hostname < 5)) {
#precaution
return 1;
}

You throw like a girl

Yesterday I tried to give Jil and Velvet some pointers on how to throw a softball. Both of them had issues turning their shoulders properly, which isn't a big deal if throwing small distances, but if you really want to throw like the boys you have to turn those shoulders. My other big tip is to get your elbow up. By the end of the practice they were both throwing much better.

I think girls can throw just as well as boys, but I think they don't get enough practice. You won't throw like a girl for very long if you keep practicing because it's just not going to go the distance you want nor have the accuracy you want until your mechanics are good. Eventually you body will train itself to put the arm, legs, and shoulders in the proper positions. So my biggest piece of advice is to keep practicing. Be sure to start slowly though, throwing a ball overhand is not a natural motion for the human arm. It is very easy to get a sore elbow or shoulder so be sure to stretch and warm up properly.

I'm looking forward to our softball season, but I hope it cools down soon, yesterday my Jeep told me the temperature was 105....

Photos From My Phone

It's been over a year since I cleaned up the photos on my phone, here are some good ones...

Colin Cowherd UVA Rant

If you haven't heard this clip and are a VT Hokie fan, check it out... UVA fans probably shouldn't listen.

Local News

Webmail.us announced today that we are now staffed in our customer care division 24 x 7 x 365. And yes, all of our customer care staff works from our Blacksburg, VA office. I'm very excited about this announcement because it is an important milestone for our company.

In somewhat related local news, it seems a lot of local communities are really pushing forward with providing wireless Internet access for all of their citizens. Check out this Roanoke Times article. I'm amazed that the smaller communities such as Bland, Hillsville, Radford, and Pulaski are actually leading the way in this area. At the same time I can't help but wonder how they plan on supporting their users.

I have a good friend who actually started a wireless ISP two years ago that primarily services the rural parts of Montgomery county plus some of Christiansburg. He's been very successful and I'm sure he's earned a lot of loyalty due to the excellent customer service he provides. I'm sure his business will continue to grow as long as he is able to continue providing great customer service.

More Debugging Tools

Here are a couple of cool debugging tools that will help you accomplish more in less time:

strace
------

strace allows you to attach to a process and watch every system call that the process executes. It is very helpful for tracking down exactly what a program is doing without having to use gdb. It is run like this (use the -f option to strace forks):

strace -tt -o log.txt -p

For an example run this:

strace -tt "uptime"

lsof
---
Another cool command is the "lsof" command. It will give you a list of open files (including sockets) for a process.

lsof -p 3387

Using GDB to Debug Linux Applications

I had to refresh my memory over the last two week on how to use the debugging application called 'gdb'. I have needed to step through both Pound and Dovecot's delivery applications to help me nail down a couple of tough problems. Debugging Dovecot's delivery application is very simple, so let's walk through it really quickly.

First, understand how the application is normally executed; normally the MTA will pipe in the email content to the deliver process, but we can fake it by using simple re-direction.

>deliver -d kevin@boebel.us < email.txt

But we want to step through the code, so we need to run it through gdb.

>gdb deliver

Before we issue the 'run' command, we need to set a breakpoint, so let's say we are trying to debug this portion of code in deliver.c (lines 540-546)

env_tz = getenv("TZ");
home = getenv("HOME");
env_clean();
if (env_tz != NULL)
env_put(t_strconcat("TZ=", env_tz, NULL));
if (home != NULL)
env_put(t_strconcat("HOME=", home, NULL));


(gdb) break deliver.c:540
Breakpoint 1 at 0x8056e72: file deliver.c, line 540.

Ok, the breakpoint is set, let's run the program, giving it the correct command line arguments:

(gdb) run -d kevin@boebel.us < email.txt
Starting program: /var/home/kminnick/dovecot/dovecot-lda-0.1/src/deliver -d kevin@boebel.us < email.txt

Breakpoint 1, main (argc=3, argv=0xbfffb3b4) at deliver.c:540
540 env_tz = getenv("TZ");

Cool, the program is now getting ready to execute this line of code, let's run some commands to get the value and type of 'home' variable both before and after the line is executed.

First, we need to skip this line of code and get to the next one:

(gdb) n
541 home = getenv("HOME");

Now we can examine the variable:

(gdb) pt home
type = const char *

(gdb) p home
$4 = 0x3f1cd8 "ä[\023"


As you can see, 'pt' prints the type of the variable, and 'p' prints the actual value. The variable is currently pointing to some garbage in memory. Let's execute the line of code and find the value of the variable after it's execution:

(gdb) n
542 env_clean();
(gdb) p home
$5 = 0xbfffdf44 "/root"

This was a very simple example of how to use gdb to debug your applications, next time we will discuss a more complex example on how to debug a program that uses forks and threads. I'm sure you can't wait... :)

Review of 2006 Goals

Today I benched 200 pounds one time. This was my goal for the entire year, so I am very happy with my progress thus far. Now I'm going for 225 by the end of the year.

Here are the goals I blogged about in January for this year and the results thus far:

1. Blog More
I'm definitely doing this, probably boring everyone a bit. :)

2. Go to the Doctor
Hmm...still six months left.

3. Get my weight to 175 lbs
I'm at 167 (I probably started the year around 163), only 8 pounds to go.

4. Oversee the creation the fastest webmail client on the market.
We are very close...

5. Become the best hitter on my baseball team.
We've only had 2 games so far, but our team captain is kicking butt. I'm hitting .375 with 2 RBI and 2 runs scored. That puts me tied for third (with my brother) on the team. We are 1-1 as a team. Mannin, our captain, is hitting .778 with a homer and 4 RBI. I have a long way to go, perhaps second best hitter on the team will be a good goal.

6. Shoot a round of golf under 110.
I've only played 3 rounds of golf this year. All at The Hill in Blacksburg. I have only had time to play 9 holes each time, but I think my scores have been fairly good, 51, 55, and 60. They are actually terrible scores, but I've made a lot of progress since last year. I think I definitely have a shot of getting a round under 110 if I ever find time to play a full 18 holes.

7. Take my family to Disney World.
Done, we had a great time.

8. Successfully manage our ever growing group of developers.
Ask Pat.

9. Bench Press 200 lbs one time
Today I did this, next up 225.

10. Find more time to spend alone with my wife.
Still working on this, but my favorite memory so far this year has to be the great time we had at Bill and Beth's wedding. We also had a great 5th year anniversary weekend. It's amazing how hard it is to find time alone when you have two kids. I'll keep working on it though.

Windows Vista Tips

Thanks to "Rick H." for posting this tips for Vista:

--------
Just a quick note to say that you can change the default search provider in IE7. Click the drop-down arrow at the edge of the search box in the upper right. Select "Find more providers" and see if your favorite is available (most are). Then click the drop-down again and choose Change Search Defaults.

For the admin popups (It's called UAP or UAC depending who you ask), run secpol.msc. Navigate to Local Policies->Security Options. Set UAC:Behavior of Elevation Prompt for Admins to "No Prompt" and UAC:Run All Admins in Approval Mode to disabled. (Use at your own risk!)
---------

I'm not sure how you figured out the UAC issue but what a life-saver that is for me. Thanks! Now if someone could figure out how to get my transparent windows back I'll be all set.

Windows Vista Beta 2 Review

After my computer decided to stop booting Windows XP Pro I decided to download and install Windows Vista Beta 2. It's been an interesting trial and error experience but I finally have it installed and running. I first had to ditch my old computer (Vista doesn't support the IDE driver that was built in 2001). I was able to install it on a brand new Dell, but I wasn't able to upgrade Windows XP Pro, I had to install from scratch. I'll probably be better off in the long run, I still hate how Dell insists on installing four million applications by default.

My first impression of Vista: "Cool Stuff". I like the visual changes they have made, I like the fading, the new transparency, the new start menu, the new explorer. I also like the gadget side bar, the new task bar, and the new task switcher. I also like IE 7, but there is no way to change the default search bar. It looks like it just hasn't been added yet, I am sure they will let me change it to Google eventually.

It comes with a few new apps like Windows Calendar and Windows Contact Manager, but they don't seem very useful. There is also some sort of collaboration app built-in, but that doesn't seem very useful either. Perhaps I need a few other users to help me test. The built-in search bar in the start menu seems to be a Google Desktop wannabe, but it appears to work rather well (I don't have a lot of data to test with yet).

I have only needed to install the following applications so far (they all work great w/ Vista):
Firefox
OpenOffice (Site doesn't work w/ IE 7 yet)
GAIM
Putty
Programmers Notepad
TurtoiseSVN
Thunderbird (for testing)

One really annoying "feature" that took two days to figure out is how to make my fonts look like crisp and clear. I was about to give up when I stumbled upon this setting:

Start->Control Panel->System and Maintenance->System->Advanced System Settings->Performance Settings->Click "Custom"->Uncheck "Smooth edges of screen fonts"

That setting has made all the difference in the world. Maybe today I can go home without a headache from my monitor.

Another issue is that I upgraded the video driver to the recommended version and I lost all of the cool transparency and window effects that I had on day one. :( I also haven't figured out how to turn off the annoying security pop-ups that open anytime I try to access an "Admin-Only" section of Windows.

I am finding the new control panel very complicated. It reminds me a lot of Perl in the fact that there are 50 ways to accomplish the same task. Navigating around the new control panel is frustrating, especially with the annoying "Admin Only, Are you really sure you want to do this?" pop-ups everywhere.

Overall, it looks like Microsoft has security as a top priority for this release. They are also finally catching up to Apple's cool windowing effects. And it won't take much more before they are caught up to Google on their desktop search capabilities. I can't think of any real compelling reason for anyone to upgrade besides the improved security features. Perhaps I'll find more in the coming days...

A Semi-Solution for My Internet Based File System Problem

After thinking about this more today, I have come up with the following solution that will almost work perfectly. First, we have an existing Subversion repository, so I created a "personal/kminnick" project for my own use. Second, we have an existing network drive called the "u" drive which can be used to store personal files. So, I installed a Windows based svn client called Tortoise SVN which is actually a shell extension that integrates directly into Windows Explorer. This basically means I can use Windows explorer to check out / check in files into Subversion, awesome stuff. I checked out my personal project onto our network drive (which is going to be available over a VPN eventually) so that nothing is stored on a local hard disk. To access the files from any location I can use a simple web browser to access the Subversion repository directly. If traveling, I could checkout all the files to my local hard disk, then check them back in whenever I have Internet access. Cool stuff. So with all of this setup and working I can:

1. All of my files are backed up automatically.
2. Have a history of every revision for all of my documents.
3. Access my files from any location that has a web browser.
4. Check out files locally for airplane trips, vacations, etc.

Not a bad solution, it was easy to setup and doesn't require a local hard disk for file storage.

Dead Computer

Today my desktop computer at work died. My dad gave me this computer about 2.5 years ago because he bought a new one, I'm guessing it was only about 2 years old when he gave it me. I was very attached to it, but all good things must come to an end. I was constantly asked by my co-workers why I never upgraded to a new machine. Well, for one, I never really “needed” a new machine. I spend most of my day inside of an SSH session, so memory, CPU, etc isn't that necessary. I'm curious though how many Word documents I'll need to re-type because I saved them to my local hard drive instead of on the network. I usually back everything up to the network or have them somewhere in my email for safekeeping, but I'm sure I missed a few. I use webmail exclusively for email, contacts, calendar, and tasks, so nothing will be lost there. I'm going to try to put in a new hard drive into my old machine and re-install Windows. But this time I'm going to try hard to move everything onto the network, the way it should be stored. In fact, I'm going to try to find a solution that will allow me to store everything I need onto the Internet.

This thought has really made the wheels turn this morning. I really need the following:

1. A web-based central repository that I can put / get files (any file type).

2. It should work like any good version control system (subversion), meaning it keeps track automatically of revision numbers, and I can view old revisions easily.

3. It would be great if I could open the files using “File, Open” from any application such as Open Office, Photoshop, etc. I shouldn't need to download the file first locally. I imagine this would require some sort of add-on to Windows Explorer. It needs to be fast though, so the plug-in should pull it down to a local temporary location, but automatically save it on the Internet when I click “File, Save”.

4. It should support sharing, so instead of having to email a document to everyone, I could simply email a note to everyone I want to share the document with and they can pull it down after entering an email / password. The link should always point to the latest revision so that timing isn't an issue. You should also be able to setup who can make modifications to the document.

5. You should have the ability to setup notifications so that you know when the document has been changed / modified.

6. It should be integrated inside of webmail.

For now, all I really need is #1. I think to solve that I'll setup a subversion server for myself and put all of my documents into subversion. I will be able to access it from home and work, the only downside is that I won't be able to open / save documents without the help of a subversion client. But at least everything will be backed up automatically and all of my documents will be accessible from the Internet.

If anyone knows of a solution that will meet the first five requirements above please let me know. I honestly haven't researched it yet.

Trac

Yesterday I setup and installed Trac for our development team. Trac is one of those tools that once a developer sees it, they gotta have it. If you haven’t heard of Trac, it allows developers to easily browse Subversion repositories. Subversion is the tool we use to track code changes. Subversion doesn’t have a good web-based GUI for browsing the repository, which is the primary reason we wanted to install Trac. Trac is easy to navigate, provides timelines, and makes browsing differences between revisions easy-easy-easy. The developers at Edgewall really have a winner with this product.

Trac also has a bug ticketing system, wiki, report generation tool, and milestone tracking tools built-in. The other tools need a bit more flushing out, especially the bug-tracker, but the SVN browser makes it worth the trouble to install. The installation process for Trac is…well…tough. I’ve installed hundreds of applications by source and RPM over the years and Trac was definitely one of the hardest I’ve ever encountered. Not only are there a ton of dependencies, but you have to be very careful that you have the right versions of each dependency. Thank goodness Google was able to help quite a bit.

I can see the benefits of Trac already paying off; just this morning, Mike sent me a link to the change-set in Trac for the next upload. Clicking on the link takes me to a web page with a description of the changes, a list of all files that were changed, and a color-coded side-by-side diff for each file….great stuff.

Thanks to Joel, one of our latest hires, for directing us to this application.

5th Wedding Anniversary

When I was in the basement today I saw a picture album from our wedding night and decided to reminiscence a little. First, it's amazing how much changes in five years. Second, it amazing how much stays the same. Velvet, Bill, Jason, and Velvet's parents all look the same today as they did five years ago. Pat, my parents, and my brother look totally different. I look like another person and wonder what Velvet was thinking...I'll bring the pictures to work if anyone is interested in looking.

To celebrate our anniversary we dined at The Summit restaurant in Christiansburg for the first time. Velvet suggested it and I heard three reviews all raving about the great food. I have to admit the food was really good (even though my steak came out a little too done). If you are looking to impress your date with some great food and a nice atmosphere I definitely recommend The Summit. :)

I'm more of a “look to the future” guy than “dwell on the past” guy; so I'll give you some insight on what will be going on five years from now, instead of what has transpired over the last five....In five years Drew will be in kindergarten and Jordan will be in 4^th grade. I'll be working on my 11^th year at Webmail.us. I'll only have 20 more years to pay on my house (haha). Maybe I'll have the driveway paved by then so I can put up a basketball goal for Jordan. Hopefully I'll still have my Jeep (gas will only be $6.00/gallon but that still beats having a car payment). Maybe my golf game will be in the 90s (OK, so I'm dreaming a little). Jordan won't remember a time when TV wasn't in high definition. Wireless Internet access will be available practically everywhere for free. Jordan will be using this laptop and think it's archaic. I might be caught up on the Sopranos. I hope Jordan can teach Drew how to play chess, read, and create his own blog. Jordan will still be in gymnastics (and will keep requesting a trip to Sonic after each session). Velvet and I will be celebrating our 10^th Wedding Anniversary.

OpenLDAP Errors

I ran across the following error in OpenLDAP two weeks ago when deleting an entry:

"Additional info: entry index delete failed"

Huh? We've been using OpenLDAP since 2003 and I've never seen this problem before, but looking at the OpenLDAP code showed it was trying to delete an index. We use the BDB backend and apparently a BDB index was corrupt. I haven't found an easy way to repair these indexes besides dumping all the data and re-importing. Maybe a BDB expert out there has some suggestions but luckily we replicate all of our data across multiple machines so I just had to copy the data (zip up the /var/openldap-data/ directory) from another machine. Googling for the error above didn't return many results so I thought I would go ahead and post my findings for the next person who searches for this error.

Harbinger Beta Available for Public Testing

A beta of the latest version of our webmail client is available to any current Webmail.us customer. Simply log in with your email address and password at http://beta.webmail.us.

You will notice only a few minor interface changes as our primary focus for this version is speed and performance. We no longer use frames and we make heavy use of AJAX and caching to improve speed. The first time you log in to the application will take a bit longer than normal because the browser is busy downloading JavaScript code that will be cached for future sessions.

Switching between folders, calendar, contacts, and tasks is significantly faster than previous versions of our application. We believe that a web based email client has significant advantages over a desktop based client but only if the web based client can match the speed and performance of the desktop client.

I've been using Harbinger for a few months now and have been really impressed with the speed thus far. We believe this is only the tip of the iceberg though and will continue to push the limits of the browser to maximize the speed of every aspect of the application.

Disney World Wrap Up

We had a great time at Disney World and I'm glad we were able to have a family vacation. I think that's the first family vacation we have taken since Drew was born. The kids did great on the plane flights, I'm so proud of them. Disney amazes me each time I go with how big they think. Nothing they do is too much, too big, too hard. They have big shows, big parks, big fireworks, and thus big crowds. They really do a great job and make you want to come back for more. We might head to Va Beach one weekend this summer, but otherwise I think we are done with family vacations for awhile. I'm glad I got to spend so much time with Drew, Jordan, and Velvet. If we do go back next year I'll have to remember these facts:

1. Go back to Magic Kingdom, Jordan loves it.
2. Don't go back to MGM Studios, it was fun, but not good enough to do two years in a row.
3. Make reservations at Cinderella Castle for dinner at least a month in advance.
4. Rent from a car rental place at the airport, not Hertz or Thrifty. (Hertz was great, but too hard to get on and off the bus.)
5. Caribe Royal is a great hotel, easy check-in, easy check-out, great pool, easy parking, but don't order the room service, very expensive and the food just wasn't that great.
6. Eat at Smokey Bones again.
7. Buy souvenirs at Wal-Mart in Orlando, nobody will know the difference and they are 1/10th the price.
8. Don't try to stay until close for the fireworks, the kids could probably make it, but I can't.
9. The "park hopper" option isn't worth it.
10. 10:30am is a perfect flight time, both ways. Connect in ATL at 1:30.
11. Mother's Day weekend is a good weekend to go because there is no baseball.
12. Jordan's favorite ride is the tram from the parking lot to the park, so ride it around a few times.
13. Jordan's favorite part of the trip is the plane ride and the pool, so take a trip to the beach with a Cinderella DVD and save some money. :)
14. Bring your own map of Orlando, the one from Hertz might as well been created by Jordan on the airplane ride down there.
15. Bring your own band-aid's so that you don't get weird looks from the first-aid people when Jordan gets a scratch on her leg.
16. Don't leave the diaper bag unattended next to the Haunted Mansion again, the next time someone might think it's a bomb and you don't want to be the one responsible for shutting down Disney World.
17. Invite some family so that you can go golfing while they take the kids to Disney World.

Disney World

Tomorrow I am packing up the wife and kids and heading to Orlando for our family vacation ... man I am old. :) Should be fun though and I'm sure Jordan will have a good time. We are heading back on Sunday.

Update on MySQL Issues

Last week I wrote about a strange issue with MySQL replication and what I thought I wouldn't see again for a long time. Well, around 1:00am this morning another MySQL server (a different cluster than previous) had the exact same issue as the first cluster. This time, my first step was to reboot the machine to see if it would fix the problem and thus not require me to spend a lot of time exporting and importing data. The reboot fixed the problem with the constant restart of the mysql server, but the replication thread failed to start. In fact, it was failing to start the slave thread on both machines. Issuing ‘start slave’ produced this in the error log of the machine with the problem:

060509 1:32:09 [Note] Slave I/O thread: connected to master 'replxxx@xxx-slave1:3306', replication started in log 'xxx-slave1-bin.004981' at position 79
060509 1:32:09 [ERROR] Error reading packet from server: Could not find first log file name in binary log index file (server_errno=1236)
060509 1:32:09 [ERROR] Got fatal error 1236: 'Could not find first log file name in binary log index file' from master when reading data from binary log
060509 1:32:09 [ERROR] Slave I/O thread exiting, read up to log 'xxx-slave1-bin.004981', position 79

This time I was determined to fix the problem without dumping data so I Googled a bit for the error and found this great post:

http://archives.neohapsis.com/archives/mysql/2004-q1/2000.html

Ah, so after creating the missing file and issuing a ‘reset slave’ the problem was fixed. On the other master I simply had to run a ‘change master to…’ command and ‘start slave’ and now replication was working in both directions again.

I think there is definitely a bug in the MySQL code somewhere, but I’m not sure what triggers it. I don’t think it’s related to the number of days a server has been up, but it might be related to the number of connections that have been opened since it was last restarted. It may just be a corrupt relay log file.

NCTC TechNite 7.0

Tonight the NewVA Corridor Technology Council (NCTC) is hosting TechNite 7.0. TechNite is self described as "the premier event honoring the NewVa Region's most promising technology companies, leaders, and entrepreneurs." Awards will be handed out in the following categories:

High Tech Leadership
Excellence in High Tech Manufacturing
Excellence in Life Sciences
Excellence in IT Development
Excellence in IT Deployment

Webmail.us is a finalist for the "NCTC Excellence in IT Development" award. Pat, our CEO, is a finalist for the "High Tech Leadership" award. It's quite an honor to be a finalist, especially considering the competition:

NCTC Excellence in IT Development Award
Webmail.us, Inc.
Phoenix Integration, Inc.
Synchrony Industrial Controls, Inc.
Integrated Imaging

NCTC High Tech Leadership Award
Margaret "Elaine" Carver, County of Roanoke
Kenneth Ferris, LUNA Innovations Incorporated
Patrick M. Matthews, Webmail.us, Inc.
Thomas A. Dingus, Virginia Tech Transportation Institute
R.J. Kirk, Third Security, LLC
Joseph W. Meredith, Virginia Tech Corporate Research Cent

Thanks go out to Trevor who played a big role in getting us nominated for these awards.

If you see Pat today be sure to wish him luck and ask him if he has that speech ready.

MySQL Replication and Heartbeat

Velvet, Marisa, Joel, Robert, and Andrew all did great in their respective 1/2 marathons and full marathons on Saturday. I certainly have a lot of respect for anyone who can endure all of the pain and agony that is involved. After we got back to the hotel I laid down for a power nap before dinner (yes, it was tiring watching all of those runners). I woke up to a phone alert letting me know a MySQL server was acting up. I log into the system and I'm seeing all sorts of errors and continuous notices that MySQL is restarting for every new connection. I've never seen this problem before, so I stop heartbeat so the other master takes over the IP. We run most of our sql servers in a master-master configuration. Master-Master means either server can immediately take over as the master if one goes down.

Thanks to heartbeat and MySQL replication, it allowed me to take my time and debug the problem without any service interruption. As for the debugging, I didn't get very far. I spent about an hour looking through log files, moving directories around (including to a ramdisk to make sure it wasn't a hard drive acting weird), etc. I was about to give the order to re-image the machine and start over when I had Bill reboot the machine because sometimes that magically fixes things. Well guess what, it did. I didn't check the uptime before the reboot, but according the other master it had been up about 284 consecutive days without a reboot. Not bad.

After the reboot fixed the restart problem, I decided to take a few safety precautions and assume the data was corrupt. So I moved the data to a backup copy, dumped everything from the master, re-imported, ran several test, and fired heartbeat back up. Everything looked good, and the whole process took about 4 or 5 hours. The restore took awhile because we have some tables that have over 9 million rows.

So my vacation in Nashville turned into a work weekend thanks to a MySQL server going crazy. Luckily we have the proper monitoring, replication, and failover systems in place, thank you Bill and team!

Big Weekend

This weekend I'm heading to Nashville via Knoxville to "watch" Velvet run in the Country Music Marathon and 1/2 Marathon (she's doing the 1/2). She's been training non-stop (I rarely awaken when her alarm goes off at 5:45am or whatever god-awful time she has it set to). I can't imagine running that far, but I can imagine that I will have fun this weekend. Our first stop is in Knoxville to go country line dancing at the world famous "Cotton Eyed Joe's". I haven't been there, but Jil says it's great so I can't wait. Based on the pictures from the website it looks like a grand ole time to me. And no, I won't be doing any dancing.

From there we head over to Nashville (first time for me) to check in at the hotel an get settled in before the big race. I'm sure Velvet will be nervous as hell (hehe). And with 20,000 participants I hope I can at least find her in the pack.

I wish Velvet, Jil, Marisa, and everyone else we are going with the best of luck...

Webmail Load Balancing Improvements

Last night we completed a “behind the scenes” project for webmail that will help us improve the scalability of our webmail server farm, improve our ability to measure the performance of our applications, and improve the overall speed up our applications. I started this project way back February and I previewed what we were doing in this post. By utilizing a piece of open source software called “Pound” we will be able to accomplish all of the goals listed above.

The primary reason for starting this project was our need to differentiate static web pages from dynamic web pages and send them to different clusters depending on the type of traffic. This would allow us to better cache the static content and get a better measurement on how well the dynamic pages are performing. We also wanted an easier way to manage all of our SSL certificates. Bill’s team (Tony and Korey) created a set of web pages that our support team can use to install and configure new SSL certificates. Mike also chipped in any necessary code changes that were need for webmail. And finally, Vinny helped me re-configure the control panel application for the new cluster of static servers.

I appreciate everyone’s help in making this project a complete success.

Tech Talk

I benched a personal best of 155lbs 10 times morning (which means I can pretty much bench my weight 8-10 times which is pretty good), but that's not what inspired me to post a blog...

Last night Me, Mike, and Pat gave a 'Tech Talk' on the Virginia Tech campus to a small group of computer science students. Basically we talked about Webmail.us, our development environment, our projects, the technologies we use, and we answered a lot of questions. I think it went well and we had a good time. One question that I did not get to answer was from a student who wanted to know what it was like to start your own company in Blacksburg because he was pondering the idea. Pat gave a great answer, but I couldn't help but think about this on the drive home.

My advice to anyone who is coming out of college and thinking about starting their own company is simple. As long as you are ready to put in long hours and are not afraid of failure, I think you should go for it. I think you will fail. But that's OK. Life is short, if you take that high paying job at IBM and Microsoft, that is the safe thing to do. It takes absolutely no courage to be that person. It will also prove to be unrewarding and boring in two years. Even if it's exciting work, you haven't learned the tough lessons that go along with starting your own company. You haven't learned how to fail, and you won't appreciate what it means to really work hard. If you are 22 and coming out of college, there is no better time to learn life's tough lessons. You have no kids, no mortgage, no obligations. It doesn't take a lot of money to start a company, you basically need an apartment and a work ethic. There is no better time than the present.

If you have an idea, a passion for that idea, and are not afraid of failing, GO FOR IT!!!

Of course, if you are afraid of failing, or you just don't have a great idea, I suggest working for a small company as a first step towards starting your own company. At a small company you will see and experience what the owners of the business are challenged by on a daily basis. You will see it's not easy. But this is great experience. After you work for a small company you will quickly realize that either that's the life you want to live or it's not. Perhaps you will find out that you do belong at a big company. Whatever the case may be, don't be afraid to take chances every once in awhile. That's my advice for all you graduates, good luck with whatever you do decide.

Product Development Releases Last 60 Days

Looks like I forgot to post this last month, but here are some of the great things we released over the last 60 days:

* Website: Printable Pages
* Control Panel: Improved Alias Management
* Control Panel: Simultaneous Logins
* Control Panel: Redesigned Login Page
* Website: Email Traffic analyzer
* Webmail: Address Book Import / Export Improvements
* API: Added 3 new functions
* Website: Redesigned Home Page
* Control Panel: Improved Availability
* Webmail: Spanish Language Improvements
* Webmail: Ability to select default language, and html font

There's plenty of stuff I can't mention too, but overall we are really busy around here and are constantly pumping out bug fixes, new features, even a beta. ;)

March Bench Press Results

Technically I have another week in March, but I don't think next week will be better than today, so I'll go ahead and post this month's progress:

185 x 3
195 x 1 (2 with a spot)

My bully spotter made me try 205, but I needed a little help from him with that, maybe next month. I definitely should be able to hit my goal of 200 next month, so I might up the ante to 225 by the end of the year. The best part of all of this is that it's been the same two people in there spotting me every Friday since I started in January and they were saying today they can really tell a difference. Hopefully the progress will continue...oh, and I weighed in at 165 today.

Google is Yahoo 2.0

Jeremy posted this awhile back, but with the recent launch of Google finance, I totally see what he is talking about. I really enjoyed Jeremy's latest comments on how Yahoo! had many of these same ideas years ago, but due to many issues they were never able to act on them.

I think Yahoo! and Microsoft are very much in the same boat on this one. All of the competition from open-source in regards to the browser, the desktop, the office application suite, the server market, and the software as a service model is very similar to the Google vs. Yahoo! fight.

I certainly hope that our commitment to keeping our product teams small and focused will prevent our software from becoming stale. I want our teams to be able to release high quality software and then immediately be able to turn around and say "How can we make this even better?"

Congratulations Bill & Beth

Last weekend Bill & Beth were married in Alexandria, VA. I have known both of them it seems like forever and they are really great friends. They certainly know how to throw a great party too and their wedding was no exception. I think everyone had a great time and I was glad so many of our friends were able to attend.

So who's the next one in our company to tie the knot? Manny? Pat? Here are my odds:

Manny 3:2
Brian 11:5 (Vegas is only a plane-ride away!!!)
Pat 2:1
Mike 4:1
Steve 5:1
Huey 8:1
Matt 10:1
Ben 10:1
Vinny 12:1
A part timer 20:1


Any takers?

My Evaluation of Reverse Proxy Software

A proxy server takes outbound traffic and makes it appear as if it comes from a single IP address. This type of proxy is useful if you have a single access point to the Internet and want multiple computers behind that access point to be able to access the Internet. This is common in an office or corporate environment.

A reverse proxy takes incoming traffic and sends it to multiple backend computers. This is useful if you run a large website and want a single IP address to actually be served by multiple machines. You can also accomplish this with a load balancer, but load balancers typically lack several key features that a reverse proxy can provide.

Let's take a look at 4 reverse proxies that are currently available:

1. Apache's httpd
2. Squid
3. Perlbal
4. Pound
5. HAProxy

All of these are very good, but depending on your requirements some may not meet your needs. I'm interested in a reverse proxy that can handle the following requirements:

1. Easy to configure
2. Small memory footprint
3. Provides load balancing
4. Supports SSL to client, but non-ssl to the backend
5. Provides High Availability
6. Can proxy traffic based on URL pattern matching
7. Supports virtual hosts

Perlbal is really a terrific piece of software, but it has a lot of features that I don't need, plus it does not have one key feature that I do need, SSL. Although there is an option to configure Perlbal with SSL, don't do it. It's too easy to break because of this problem reported on their listserve. It appears this is not a problem with the Perlbal code, but a problem with the underlying libraries that Perlbal uses. I could stick stunnel in front of Perlbal and I think that would work great, but you have to setup one instance of stunnel per SSL site and that wouldn't work for me. Perlbal uses epoll, which would provide a big performance boost. Perbal doesn't support sticking users to backends (for session support), but that is not a requirement for me.

Squid has been around for a long time, but I think it's too bulky. First, the configuration is too hard. Second, it's primary goal is to serve as a proxy, but I do not need this, just the reverse proxy. Squid also does a lot of caching, I don't need this either, I would prefer to have something that never touches disk. Also, to my knowledge, Squid does not support high availability (checks to make sure a backend is not down).

Apache's latest release (2.2) of it's httpd server does have better support for load balancing and high availability than previous versions, but it's still too bulky for me. One thing that kills me about Apache is how hard it is to configure mod_rewrite. I actually was able to setup Apache 2.2 as a reverse proxy and it was working perfectly, but I realized that this was going to be too hard to maintain on a large scale system. It's great that Apache is so configurable and can handle just about any task, but that can also be looked at as a big negative too.

Bill showed me HAProxy and it looks great, very configurable. But it requires Apache for SSL support, and I don't think you can proxy traffic based on URL.

The final piece of software that I looked at as a possible reverse proxy that would meet all of my requirements was Pound. I had a little trouble compiling Pound at first on RHES3, but I finally figured out the right compiler flags. On RHES4 it compiled with no problems. Besides that Pound seems to work great. It's easily configurable, is designed to accomplish only one task, and is small in size.

So there you have it, my evaluation of reverse proxy software. If I missed anything or I am just dead wrong, feel free to leave a comment. I'll update this post if anything is inaccurate.

February Bench Press Results

Another month has come and gone and I know you are dying to know how much I benched pressed today, so here ya go:

155 x 8 times
175 x 3 times
185 x 2 times
190 x 1 time

190! Holy cow, only 10 more pounds to reach my year end goal! Not bad considering I weighed in at 163lbs today. I haven't gained or lost any weight since I started working out, that might be more amazing than the 190lbs I lifted today.

Cryptic Apache Error

Each night our webmail servers automatically restart httpd to clean up memory and start fresh. Last night one of our servers hit this error upon restart:

[crit] (28)No space left on device: mod_rewrite: could not create rewrite_log_lock Configuration Failed

I had never seen this error before and had no idea what it meant, but Google helped me out:

http://www.goldfisch.at/knowledge/224

After some testing, I found that Apache leaks a couple of semaphores after each restart and eventually w/o a server reboot will run out of resources. So I changed our restart script to clean up these semaphores each night.

ipcs -s grep nobody perl -e 'while () { @a=split(/\s+/); print `ipcrm sem $a[1]`}'

In case you are wondering, it's not a big deal if a webmail server doesn't restart properly. Our load balancers quickly take a down server out of rotation so that customers are not affected. Thank you keepalived.

Development Releases Last 30 Days

Over the last 30 days we have released some really great additions to our platform. Here is the high level list I presented at our management meeting on Monday:

* Webmail Lite w/ Calendar / AJAX
* Control Panel Security Additions (IP Based Restrictions)
* 20+ Webmail Bug Fixes
* Webmail Import / Export Enhancements
* Company Directory
* Website Search
* 30 Day Free Trial

You can read the details on our blog.

OpenLDAP Replication

Today I setup a new LDAP slave machine, the process is fairly straight-forward, but a good thing to know about replication is how to do it without shutting down the master for an extended period of time. Let's take a look at how to do this:

First, let's verify we have a few things:

1. The LDAP ssl port is open from the master to the slave:
ldap-master> telnet ldap-slave 636

2. Make sure the slave has a valid SSL certificate. If you are using a self-signed cert you will need to make sure the master has the slave .crt file installed. If you have multiple slaves, you need to create a combined.crt file and concat all of your certs into one file. Then add an entry to the master ldap.conf (not slapd.conf) like so:

TLS_CACERT /usr/etc/openldap/certs/combined.crt

OK, so now we need to get all of the data from the master to the slave without losing anything during the transfer. Normally this would require the master to be taken offline for some period of time while you zip up the data files and scp them over. But, luckily I have another slave that is not being used for queries, so I can follow these steps to get the data moved without having to shut down the master for an extended period of time:

0. Stop the new slave.
1. Add the necessary

replica uri=ldaps...

lines to the master slapd.conf file.

2. Restart the master.
3. Stop the old slave.
4. Tar/Zip the data, scp to new slave.
5. Start the new slave, verify replication works to the new slave.
6. Bring back online the old slave, verify replication is still working.

That's it, really simply. One trick you should know about replication is that if you have multiple LDAP databases being replicated, slurpd will not replicate the second database to a machine with the same hostname as the first database. I have to create a fake host entry in /etc/hosts so that slurpd doesn't get confused. It took me awhile to figure this one out.

Bench Press Update

I've been working out consistently this month, pretty amazing, but I found that going in the morning is the only way I can do it. Today was the last chest workout for the month, so I thought I would log my progress here.

Today I did:
135 x 12 reps
155 x 6 reps
165 x 3 reps

165 is a long ways from 200, but I think I made good progress this first month. I started the month doing 115 x 12 times.

New Webmail Lite and Control Panel Features

We just finished releasing the next version of Webmail Lite. This version has a calendar, task list, and uses AJAX for speed improvements.

We also added a feature to our control panel that allows administrators to use IP addresses to restrict logins.

Read all about these these enhancements on our company blog.

2006 Goals

It's that time of year for resolution and goal setting, so here is my list for 2006 (in no particular order):

1. Blog More
I'd like to use this blog to record interesting things in my personal life and to document any research or technical details about the work I do at Webmail.us.

2. Go to the Doctor
Since we have great health coverage at Webmail.us, I better start using it. My dad has had a lot of health problems recently and let's just say it's been a wake up call for me.

3. Get my weight to 175 lbs
I think it's a perfect weight for me, and this goal will force me to attend the gym regularly.

4. Oversee the creation the fastest webmail client on the market.
Speed, speed, and more speed. We are working hard on this problem and it is the focus of our development efforts this year. I'm relying on the great developers we have to help me accomplish this goal.

5. Become the best hitter on my baseball team.
The past two years I have been the second best hitter on my team. This year I really want to take over that top spot.

6. Shoot a round of golf under 110.
I would say under 100, but that might be a bit unrealistic since I never have time to practice.

7. Take my family to Disney World.
Velvet and I work so much that we rarely take family vacations, so I would like to take the kids to Disney World this year. I know Jordan will love it almost as much as Velvet. :)

8. Successfully manage our ever growing group of developers.
I think this will be my biggest challenge this year. It's also the challenge I'm looking forward to the most.

9. Bench Press 200 lbs one time
I use to be able to do this, then I stopped working out. I'm probably at 165 right now (I use to do this 8 times w/o problems), but I honestly haven't tried a low number of reps in a long time.

10. Find more time to spend alone with my wife.
For obvious reasons... ;)

Understanding memcached memory management

memcached is a very fast program that can be used to cache data stored in a database for quick access and dramatically reduce the load on your database servers. It's one of those "must haves" if you develop a highly trafficked website that has a database backend. Understanding the internals of memcached may be a bit confusing at first, the details aren't documented well, and there are some tools that aren't that easy to understand what they do. I'll try to explain here by example.

Let's say you have a memcached setup and have allocated 256MB of RAM to the program. memcached works hard to avoid memory fragmentation, it does this by taking the memory and dividing it into classes, numbered 6 through 17 (don't ask). As you insert into memcached, it will take the size of the data you are inserting and inserts it into the class that is big enough to handle the data. See the chart below. The classes in the below memcached are of size 128B - 16KB. Ignore the Max_age column for now. Each class that has data is allocated at least one 1MB page. So, class 7 has 4 1MB pages, each data item in that class is using 128B and the class is full, so there are 32767 items in the class. (If you insert an item into memcached of say 100B, the class uses all 128B to avoid fragmentation.)

Now if we try adding to this cache, we won't necessarily get an "out of memory" error because memcached uses a LRU (least recently used) algorithm to kick out the item from the page that has been accessed the least. But, if we attempt to insert an item of say 15B, we will get an "out of memory" error because class 6 (the class would be of size 64B) has no items, and thus no pages have been assigned. But, all pages are currently being used, and memcached cannot allocate a 1MB page to class 6. In order to make this happen we would have to move a page from another class to class 6.

Let's examine how this is done:

First, let's get a snapshot of our current allocations:

[root@sql-slave1 bin]# ./memcached-tool localhost
# Item_Size Max_age 1MB_pages Full?
6 B 0 s 0 yes
7 128 B 742942 s 4 yes
8 256 B1175448 s 2 yes
9 512 B1279471 s 1 yes
10 1 kB6493931 s 1 yes
11 2 kB 434740 s 71 no
12 4 kB 917212 s 140 yes
13 8 kB 794005 s 30 no
14 16 kB1373021 s 7 yes
15 B 0 s 0 yes
16 B 0 s 0 yes
17 B 0 s 0 yes

Now, let's go ahead and try to insert 2 bytes of data:

[root@sql-slave1 bin]# telnet localhost 11211
Trying 127.0.0.1...
Connected to localhost (127.0.0.1).
Escape character is '^]'.
set key1 0 0 2
SERVER_ERROR out of memory

OK, just like I thought, memcached can't allocate a 1MB page, so let's help it out:

[root@sql-slave1 bin]# ./memcached-tool localhost move 14 6
Success.

Now, let's look at our pages again, as you can see 1 page has been moved from class 14 to class 6. The LRU items in class 14 go bye-bye.

[root@sql-slave1 bin]# ./memcached-tool localhost
# Item_Size Max_age 1MB_pages Full?
6 64 B 0 s 1 no
7 128 B 743011 s 4 yes
8 256 B1174812 s 2 yes
9 512 B1279618 s 1 yes
10 1 kB6494078 s 1 yes
11 2 kB 434887 s 71 no
12 4 kB 917208 s 140 yes
13 8 kB 794152 s 30 no
14 16 kB1373168 s 6 yes
15 B 0 s 0 yes
16 B 0 s 0 yes
17 B 0 s 0 yes

Let's insert our data again:

[root@sql-slave1 bin]# telnet localhost 11211
Trying 127.0.0.1...
Connected to localhost (127.0.0.1).
Escape character is '^]'.
set key1 0 0 2
aa
STORED
quit
Connection closed by foreign host.

Cool, it worked. Looking at our stats shows the item is aging:

[root@sql-slave1 bin]# ./memcached-tool localhost
# Item_Size Max_age 1MB_pages Full?
6 64 B 14 s 1 no
7 128 B 742970 s 4 yes
8 256 B1174852 s 2 yes
9 512 B1279658 s 1 yes
10 1 kB6494118 s 1 yes
11 2 kB 434918 s 71 yes
12 4 kB 917228 s 140 yes
13 8 kB 794192 s 30 no
14 16 kB1373208 s 6 yes
15 B 0 s 0 yes
16 B 0 s 0 yes
17 B 0 s 0 yes

Conclusions, you better keep an eye on the data in your memcached, and knowing what type of data you are storing will help you maintain it. Writing scripts to automate the moving around of pages based on data should be fairly easy.

Also, there are some little known commands that I use a lot:
stats
stats items
stats slabs
stats sizes

memcached is a great piece of software, many thanks to the Danga team.

Newsworthy

Looks like my wish came true:

http://www.hokiesports.com/football/recaps/2006016aaa.html

and Google has bundled some common software and made it easy to download and install, I haven't tried it, but it looks pretty useful:

http://pack.google.com/

Maybe I'll wait for them to release their operating system. I'm very skeptical this will happen anytime soon. It would have to be built on some sort of Unix flavor and the bottom line is that most of the applications they have built don't have Unix versions. Of course, it is Google, so I'm sure it's on their drawing board somewhere.

529 Savings Plans

I just finished signing Andrew up for a 529 savings plan. For those of you who don't know, 529 plans are a great way to save for college. They invest in stocks, bonds, etc, just like normal mutual funds, but the money has to be used for college. The funds have to be sponsored by a state, so each state has a fund. My financial advisor, Clark Howard, has some great tips about 529 plans, including his "Dean's List" and "Honor Roll List" of states with the best plans. One piece of advice that I don't listen to, even though I agree with is

"But first remember my rule that you don't save a penny for college unless you are already saving the maximum you can for your own retirement. College can be paid for with grants, loans, scholarships and work. Retirement happens only if you have saved the dough."

I don't think the whole "saving the maximum" thing will happen until Jordan is about 33, a little too late to start saving for college. I guess I'll just have to work a few more years, but I like working so no big deal. When I enrolled Jordan, I picked the Missouri state plan because it was on his Honor Roll list ( I don't think he had a Dean's List 4 years ago) and you could sign up for the plan online. This time around I wanted to chose one of his Dean's List plans so I went with the New York plan. The plan is managed by Vangaurd and you can sign up online.

Both the Missouri and New York plans let you invest in "age based" options, meaning the plans become less aggressive as the beneficiary gets older. This is exactly what I want plus it requires zero maintenance on my end, a big-time bonus because I hardly have time to pay attention to it.
I looked into the Virginia 529 plan too (my home state), but their so-called website looks like something a 2 year old put together with Microsoft Frontpage '98.

Va Tech Football 2005 Year in Review

The Virginia Tech football season has come to end and it certainly was a season filled with ups and downs. I enjoy attending the games and have seen the football program grow from being just awful to exceptionally good year after year. Becoming a winning program does, apparently, come at a cost. If I had one wish for next year's team, I wish that Frank Beamer would get some guts and bench players that cause more trouble than what they are worth. Some players just behave in a manner that does not reflect well on my alma-mater. I guess discipline and sportsmanship are far less important than winning. I suspect Frank has so much riding on winning these days that he can't afford to bench players that are obviously a poor reflection on the entire program. I completely understand his position, and our program is certainly not the first program to have such issues.

Perhaps Va Tech can start offering a new class for all incoming recruits "How not to look like an immature spoiled kid on national television 101". I guess if you want a squeaky clean program, attend Vanderbilt...Which bowl did they go to this year? And when is my season ticket application due? :(

Update: Since writing this post (I saved it as a draft yesterday), the athletic department at Va Tech released this statement.

Java Schools

Joel has a great blog post about "Java Schools", it's a good read and I agree with every bit of it. The test we give programmers on their first interview has questions regarding both pointers and recursion. They are two of my favorite questions, most people think they are too easy, but the fact is if you solve the recursion problem using a "while loop" it tells me a lot. If you solve it using binary operations, you're hired. ;)

I don't use a test as the determining factor for any hire, but I like to see how people go about solving problems. Some struggle, some work fast, some love it, some hate it, some even give up and leave without even trying. If you can't handle 6 simple problem solving questions though, chances are you can't handle developing software applications that are critical to the success of tens of thousands of businesses.

MySQL Replication

This morning I setup MySQL replication for two of our user / spam databases that we use internally. This is about the 10th time I have done this, so I don't even have to look up the instructions anymore, but I have to admit that the MySQL team has made the replication process ever so easy. We setup our databases in a master-master configuration with multiple slaves. Slaves can span geographic locations, for instance, from Dulles, Va to Blacksburg, Va, some 300+ miles apart. It's a very nice setup and I always have so much fun setting it up, I don't know why, but for some reason that type of work is fun for me, especially when it works the very first time I try it (and the last two times it has worked right off the bat). Thinking back to the first time I setup MySQL replication it took a few tries to get it right, and so if you are having trouble getting it to work, here are a few suggestions:

1. Read High Performance MySQL by Jeremy Zawodny
2. Turn on logging (vi /etc/my.cnf, log=/var/lib/mysql/log.txt)
3. Make sure every server has a unique id (server-id = 1)
4. I had a pound symbol (#) as part of the password, for some reason this didn't work.
5. Make sure you can telnet to the mysql port on each machine.
6. Make sure the username/passwords are correct.
7. The "show slave status" and "show master status" commands are very helpful.
8. Only replicate the databases you need, make sure there is not a lot of unnecessary traffic.
9. Testing is easy to do, so do it.
10. The /etc/my.cnf file is not used after the first reload of mysql. You need to run the 'change master to' command from that point forward.

Jeremy's book was very helpful, so I highly recommend it for the first-timer.

The best tip I can give is to setup a slave server that does nothing but replicate from the master. This is important because some day you will want to setup another slave without bringing down a live master. This is nearly impossible if you are using all of your machines for queries. To setup a new slave without any downtime, do the following:

On the slave that is doing nothing but replication:

1. Run 'stop slave' from mysql prompt
2. Run 'show slave status\G' from the mysql prompt, record the following two lines:
Master_Log_File: rss-master1-bin.000082
Read_Master_Log_Pos: 993448778
3. Run a "mysqldump" to get all of the data to a text file.

On the brand new slave, do the following:
1. Setup the databases, users, and import the exported data.
2. Run the following command from mysql

CHANGE MASTER TO
MASTER_HOST='master-server',
MASTER_USER='repl',
MASTER_PASSWORD='xxxxxxxx',
MASTER_LOG_FILE='master-server-bin.000082',
MASTER_LOG_POS=934159359;

3. Run 'start slave' from mysql (on both slaves).
4. If you have logging turned on, you will see all of the queries coming through since you stopped the slave process on the first slave machine.

That's it, it should work fine, but testing is encouraged. :)