We’re having a discussion with a new client about how to scale the database back end. Really, we’re not too worried about performance — we don’t see that being a limiting factor at this point. What we absolutely need is reliability. If the database server goes down, the site goes down. It’s a single point of failure. So how does one stop it being a single point of failure with a Rails application?

We’re currently using MySQL and, before I got involved in the project, they were trying out MySQL replication and clustering. And having very little success. I haven’t actually looked at the setup in any detail yet, so I don’t know what’s going wrong, but I’m wondering if these are the best options to persevere with?

I’m not particularly wedded to MySQL, but we don’t particularly want to go down the route of Oracle if we can avoid it. (Mostly because it sounds big and scary and the only places I’ve …

Sometimes you're so involved in the details, you forget the bigger picture--until something reminds you again.

Today I realized that our payment gateway processor (who shall remain nameless) does not support the Discover card correctly. When you authorize a credit card charge, the credit card (Visa, MasterCard, American Express, or Discover) performs an "address verification service" or AVS, to guard against fraud. This means matching the billing address on file of the cardholder against the address provided for the transaction. If the address and the zip code both match, AVS returns a "Y". If only the zip code matches, AVS returns a "Z". If neither matches, you get an "N", which in turn causes the gateway processor to reject the credit card.

It turns out that Discover returns an "X" if there is an exact match between the address provided and the billing address on file. Since Visa, MasterCard, and American Express do …

Well, it’s later than I’d wanted, but it’s here! Be sure to download it right away, as the news is time-sensitive (the rest of the segments are relatively timeless).

The feature of this podcast is “Backups, Backups Everywhere”. The show notes can be seen at:

Check out the show notes at:
http://technocation.org/index.php?option=com_content&task=view&id=18&Itemid=29

The podcast can be played directly at:
http://technocation.org/index.php?option=com_content&task=view&id=19&Itemid=29

Subscribe to the podcast at:
pcast://technocation.org/index2.php?option=com_podcast&feed=RSS2.0&no_html=1

OurSQL: The …

• Ruby on Rails Plugins - Agile Web Development Ruby on Rails Plugin directory. (tags: rails plugins ruby development rubyonrails)
• Test-First Development :: PeepCode Ruby on Rails Screencasts A new screencast from Topfunky on Test- (or Behaviour-) Driven Development with Rails. Hopefully there shouldn’t be much new content here for me, but I guarantee I pick up something interesting. (tags: …

MySQL AB today announced that Tellme Networks, Inc is benefiting from significantly reduced database-related costs by using the open source MySQL database. Tellme employs MySQL to power mission-critical enterprise and carrier applications on the world?s largest open platform for the phone.

9 years after X'97 I ended up at TUE again this time for T-Dose.org A whole different sight, last time there were a couple of hundreds geeks sitting around writing demos, compsing music etc this time almost 10 years later we just abused the auditoria and were sitting there quietly listening to people giving talks, one thing was sure.. this time I booked a hotel, no need to sleep under a table or try to sleep in my car on the parking lot :)

Altough I had no other plans this weekend I still was late for the keynote and most of the morning talks. The first talk I ran into was by "Martin WebHuis", he was talking about migrating away from Exchange to openXchange, somehow I didn't expect this kind of talk at this event, but then again I didn't expect a using Samba talk at last year's LinuxKongress in Hamburg. He mentionned openITIL which I should take a closer look at .. some day ..

What strikes me is that we in Belgium have the idea that …

With great pleasure I today noticed that we finally have daily snapshot builds of all current versions of the MySQL server source trees (4.0-5.1) available on http://snaps.mysql.com/ again. Unfortunately this has not been the case for quite some time - a few issues that caused the 5.0 and 5.1 snapshot builds/tests to fail have now finally been resolved. I also made some changes to the build script that performs the snapshot builds, e.g. to use the Perl-based test suite where available instead of the Shell-based one, which is going to phased out in 5.1 anyway. Let's hope that from now on the snapshot builds will now be available more frequently again, I'll keep a close eye on that.

Sorry for the long delay, I know that some of you depend on receiving updated source tarball snapshots on a regular basis instead of having to use the free BK client...

With great pleasure I today noticed that we finally have daily snapshot builds of all current versions of the MySQL server source trees (4.0-5.1) available on http://snaps.mysql.com/ again. Unfortunately this has not been the case for quite some time - a few issues that caused the 5.0 and 5.1 snapshot builds/tests to fail have now finally been resolved. I also made some changes to the build script that performs the snapshot builds, e.g. to use the Perl-based test suite where available instead of the Shell-based one, which is going to phased out in 5.1 anyway. Let's hope that from now on the snapshot builds will now be available more frequently again, I'll keep a close eye on that.

Sorry for the long delay, I know that some of you depend on receiving updated source tarball snapshots on a regular basis instead of having to use the free BK client...

Finally figured out what was causing the lack of performance on the
partitions with the query. The use of FORCE INDEX was causing the slowdown
with the partitioned table. Once I removed the clause, the query ran in 1
minute 19 seconds, which is more in line with expectations. Sorry for any
inconvenience!

Linux and Unixes have excellent metric of system load called "loadavg". In fact load average is is 3 numbers which correspond to "load average" calculated for one five and 15 minutes. It is computed as exponential moving average so most recent load have more weight in the value than old one.

What does Load Average corresponds to ? At least on Linux it is number of processes which are in "running" state or in "uninterruptable sleep" state which typically corresponds to disk IO. You can also map LoadAvg to VMSTAT output - it is something like moving average of sum of "r" and "b" columns from VMSTAT.

Obviously minimum value for LoadAvg is zero which corresponds to completely idle system, and there is no maximum

First thing to understand about LoadAvg it does not really tell you if it is CPU bound load or IO bound load. For example if you have LoadAvg of 10 it may mean there are 10 processes/threads actively consuming CPU or it …