Gojira est mort... Vive Gojira.

Yesterday morning, I discovered that the machine this site used to run on had quietly gone belly-up sometime during the night - it seems that the system drive and the disk controller ate themselves, or ate each other, or some such thing. Fortunately, the drive that all the site files and databases lived on survived. So after shuffling some bits around and a fun-filled day installing the operating system, Apache, MySQL, PHP, etc. and recreating all the lost configs, Gojira lives again. The old Gojira was an 800 MHz Celeron with 384 MB memory, and the new one is only a 433 MHz Pentium-3, but it's got 768 MB RAM, and actually runs loads faster. I've also got my eye on a couple of machines at the local hockshop that are faster, and with the spare RAM from the Gojira That Was, I could come close to matching my deaktop (a 2.8 GHz P-4 with a gig of memory) for less than AU$500, I think. In any case, I've come to the …

Gojira est mort... Vive Gojira.

Yesterday morning, I discovered that the machine this site used to run on had quietly gone belly-up sometime during the night - it seems that the system drive and the disk controller ate themselves, or ate each other, or some such thing. Fortunately, the drive that all the site files and databases lived on survived. So after shuffling some bits around and a fun-filled day installing the operating system, Apache, MySQL, PHP, etc. and recreating all the lost configs, Gojira lives again. The old Gojira was an 800 MHz Celeron with 384 MB memory, and the new one is only a 433 MHz Pentium-3, but it's got 768 MB RAM, and actually runs loads faster. I've also got my eye on a couple of machines at the local hockshop that are faster, and with the spare RAM from the Gojira That Was, I could come close to matching my deaktop (a 2.8 GHz P-4 with a gig of memory) for less than AU$500, I think. In any case, I've come to the …

The Linux Format magazine is having it’s annual reader awards in a number of categories.

These include (I’ve include my picks after each category):

• Best Desktop Application (Gnome)
• Best Distro (CentOS 4.2)
• Best Development Tool (Eclipse 3.1)
• Best Hardware Support (NVidia for my Dell Laptop)
• Free Software Project of the Year (Gaim)
• Embedded Linux Award
• Best Server Software (MySQL 5.0)
• Best Office Software (Open Office …

Total cost of ownership (TCO) for relational databases is, in my opinion, comprised of 3 parts. License cost, support costs and the people costs of maintaining and managing the database.

The first component, database license cost, is a one-time fee paid to the vendor for the right to use the database on one processor. The license is generally perpetual, so youll need to work this out over the expected life time of the application or the time you’re going to continue to use database software. In some cases, if this money has already been spent, you might not include it in the TCO estimates.

Second, the ongoing support costs must be included. For Informix, DB2 and Oracle the support cost is usually a percentage (typically between 15% and 23%) of the initial license purchase price. In my experience it can range between $5k and $10K per CPU depending on which database you’re using and the initial purchase price for the licenses. …

David Logan (HP) and Greg Lehey (MySQL) have started a MySQL User Group in Adelaide, South Australia.

The first meeting will be on Thursday, 2 Feb 2006, 7pm at HP House, 148 Frome St., Adelaide. Special guest speaker is Brian Aker, MySQL Director of Architecture! (Brian will also be visiting LinuxConfAU 2006 in Dunedin, and the MySQL User Group in Brisbane).

Sign up (free) and RSVP for the Adelaide event at http://mysql.meetup.com/139/.

David Logan (HP) and Greg Lehey (MySQL) have started a MySQL User Group in Adelaide, South Australia.

The first meeting will be on Thursday, 2 Feb 2006, 7pm at HP House, 148 Frome St., Adelaide. Special guest speaker is Brian Aker, MySQL Director of Architecture! (Brian will also be visiting LinuxConfAU 2006 in Dunedin, and the MySQL User Group in Brisbane).

Sign up (free) and RSVP for the Adelaide event at http://mysql.meetup.com/139/.

With release 2.7.0 the popular database web front-end phpMyAdmin starts to adopt more and more of the new features of MySQL 5.0. But there are still some hurdles on the way. If you use VIEWs on huge tables, phpMyAdmin can for example easily lock up your database.

The problem is, that phpMyAdmin tries to get an exact row count for MySQL VIEWs in many places. It achieves this by executing a SELECT COUNT(*) FROM view_name against the VIEW. This happens for example on the database overview page (that shows up by default whenever you select a database from the menu) or even for the list of tables in the left frame. Other database administration tools that use a similar strategy as phpMyAdmin could also be affected.

The disastrous consequences of the row count

This is of course a very dangerous feature for complex …

What do y'all think of the idea of a 3d printer? Aaron uses one at work. They call them "C&C machines." I don't recall what C&C stands for, but it basically takes a 3d model from a computer and cuts it into a piece of metal. I've been thinking about building such a beast.

Since there are a lot of data related to such a task, I figured I could get mysql and maxdb involved. Feedback?

What do y'all think of the idea of a 3d printer? Aaron uses one at work. They call them "C&C machines." I don't recall what C&C stands for, but it basically takes a 3d model from a computer and cuts it into a piece of metal. I've been thinking about building such a beast.

Since there are a lot of data related to such a task, I figured I could get mysql and maxdb involved. Feedback?

One benchmark I have always wanted to see performed is a comparison of SATA vs SCSI as it related to MySQL server performance. While I am pretty sure SCSI would come out on top in most comparisons, I always thought it would be interesting to see how wide the performance gap is compared to the difference in costs.

Seeing as I lack the resources to do such a comparison, I have been content to sit and wait for such a comparison to be made, and gently nudged some of the people who may have the resources to do such a comparison.

In the meantime, here’s a couple of pages from a StorageReview article on Western Digital’s new 150Gb Raptor:

Multi-User Performance - Compares the IO results for the Raptor vs 10K and 15K SCSI drives in a multi-user environment.

…