Fun with Monitoring - Monit and Munin

At work we've been having some good fun w/ monitoring packages as work.  Recently testing and then installing Monit and Munin on a variety of servers for various reasons.  If you need to collect data on the performance of a server(munin) and have some degree of alerting and automatic watching and handling of processes (monit) check them out.  They are both very flexible and useful tools that are easier to install that most others in their class.

Tim O'Reilley gives a nod to those who care about Scale and Infrastructure

In a post at oreilley.net Tim O'Reilly give a nice tip of the hat to the folks that slave away trying to shared knowledge about building web sites that are available, scalable, reliable, and performant.  He mentions a new book that just made their "internet best seller" list recently.  The book was "High Performance Web Sites" by Steve Souder.
 
Since I read the book in its pre-release PDF form I'll do a little mini-review here.  I can pretty safely say that if you are a programmer or administrator embroiled in building web sites that you would like to actually work if they get popular you should read this book.  I'd also recommend it for managers of development teams.  The way the book is broken down is easily digestible and good for getting ideas across to your team.  In summary, buy it and read it.  It's not revolutionary in most cases but it's just good sense all around and it's the stuff you just scratch your head about because you figured your devs and admins just knew this stuff anyway.  But, they just might not.

Oh, there is another link out in the article to the "High Performance MySQL" book.  Don't buy that one.  It's in the process of being re-written right now and is SERIOUSLY in need of such a thing.  It's a great book, but the content is just a bit to dated for my liking.  Here's the blog where you can follow the re-writing of that book.  They are even taking feedback to incorporate.  Pretty cool and shows you publishing really has changed a quite a bit in the last few years.  Here are some of the recent posts to that blog about the writing of the new edition.

http://www.xaprb.com/blog/2007/10/07/high-performance-mysql-second-edition-query-performance-optimization/
http://www.xaprb.com/blog/2007/10/05/high-performance-mysql-second-edition-advanced-sql-functionality/
http://www.xaprb.com/blog/2007/10/02/progress-on-high-performance-mysql-backup-and-recovery-chapter/

More Information about Vertica Webinar w/ Demo

I got a chuckle from an email I received recently for an invitation to a Webinar tomorrow.  I would like to go to the webinar so this post is no ill reflection on the actual event itself.  Here is the excerpt from the email.

"Getting 30-year old OLTP database technology to handle today's data warehouse and business intelligence (BI) workloads, is like trying to put a square peg in a round hole. As a result, DBAs are required to force databases to perform unnatural acts at high cost, or replace them with expensive and proprietary data warehouse hardware."

I believe that I can say with great confidence that I don't want to see most of the DBA's I've ever known "being forced to perform unnatural acts at high cost..."

Thanks for the chuckle Vertica.  I find the promise of the Vertica technology interesting although I'm less interested in it for data warehousing than I am for large scale web application possibilities.   If you would like to follow up on the webinar, here is the information.  

Good Things Come in Small Packages
The advantages of compression in column databases

What if your database could store 20TB of user data in just 2TB of disk space? Not only would that save you hundreds of thousands of dollars on storage hardware, it can help boost database query performance by over 50x.

Attend this complimentary Webinar to hear MIT Associate Professor Dr. Sam Madden describe how column-oriented relational databases are able to aggressively compress data and as a result, revolutionize the performance and economics of data warehouses and data marts, so that more people can query more data faster and more affordably.

Date: Wednesday, October 10
Time: 12:00PM Eastern time, 9AM Pacific
Duration: 1 hour
Featured presenter: Dr. Sam Madden, MIT Computer Science Professor and Vertica Advisor

Agenda:

Column-oriented RDBMS architecture overview
How the columnar Vertica Database is able to:
Compress data by 90%
Query it in compressed form
Use compression to enable other DBMS break-throughs such as support for grid computing and ad-hoc query-handling
Benefits of compression:
50x-200x faster query performance
70% lower storage costs and greater resource utilization
Handle ad-hoc queries without special tuning
Customer benchmarks: compression in practice in Call Detail Record (CDR) databases
Live demonstration of Vertica compression and query performance
Getting 30-year old OLTP database technology to handle today's data warehouse and business intelligence (BI) workloads, is like trying to put a square peg in a round hole. As a result, DBAs are required to force databases to perform unnatural acts at high cost, or replace them with expensive and proprietary data warehouse hardware.

Register Today!

OSCON 2007 Presentation Roundup

Even though this one wrapped up a few months ago I just found the full list of presentations online to day at the OSCON site.  So, this is probably pretty old news since OSCON ended in July.  But, since I do try to collect my thoughts here I wanted to add it the site.

http://conferences.oreillynet.com/pub/w/58/presentations.html

There is a LOT of great information contained in all those various slideshows.  So, enjoy!