One needs to be very careful before deleting files as once deleted the files are gone for good. A friend of mine today called as MySQL was failing to start on his server. Earlier when "cleaning" his server, he had deleted binary log files without fully understanding their role. When I logged in to his server, I noticed the following errors:

060716 17:21:53  mysqld started
060716 17:21:53  InnoDB: Started; log sequence number 0 1736000119
/usr/sbin/mysqld: File './mysql-bin.000002' not found (Errcode: 2)
060716 17:21:53 [ERROR] Failed to open log (file './mysql-bin.000002', errno 2)
060716 17:21:53 [ERROR] Could not open log file
060716 17:21:53 [ERROR] Can't init tc log
060716 17:21:53 [ERROR] Aborting

060716 17:21:53  InnoDB: Starting shutdown...
060716 17:21:55  InnoDB: Shutdown completed; log sequence number 0 1736000119
060716 17:21:55 [Note] /usr/sbin/mysqld: Shutdown complete
 …

Hi, what follows are my personal thoughts after reading Dmitry Yemanov's comments to a great poll posted at Firebirdnews about the features people would like to see implemented.
Quote:

Well, since Intel and AMD are massively going for multicore architectures on the desktop and SMP servers are more and more common in datacenters I wouldn't place this in a whishlist, …

I fixed a serious BC issue that I introduced in the last releases of the MySQL drivers, that would cause issues with servers that do not support InnoDB. I switched to forcing the InnoDB table handler. The fix is to simply not force any table handler and use the server default. There were also some minor bug fixes in several of the other drivers. More importantly however I am pretty happy with the nested transaction API as well as improvements to the debug callback mechanism.

I have mentioned the nested transaction support previously. Since then we have added support for SAVEPOINTs. This feature allows you to define a savepoint within a transaction to which you can rollback to, instead of rolling back the entire transaction. By default the nested transaction API will automatically use SAVEPOINTs to give really tight control in an entirely transparent fashion.

I have also …

I've been looking at a problem with our memory cache system at work for the last week and the problem finally clicked into place.

At work I developed this really awesome benchmarking library which exports statistics from various subsystems in our robot. It works really well - too well sometimes.

Its been telling me that one of my key caches is only functioning at 20% efficiency.

This is a memcached cache which means its broken out into many small servers running 256M or 512M or memory (really whatever they can spare).

Each of these servers was reporting that they are only 70% full so therefore I don't need to add any more memory right?

Wrong. I was look at the number of stored bytes vs the total maximum amount of bytes allocated to memcached.

The problem is …

A couple of days before Wednesday’s European Commission hearing, I learned that Managing Intellectual Property magazine, the leading international magazine for IP owners which has more than 10,000 readers around the globe, renominated me to its annual list of the “top 50 most influential persons in intellectual property”.

The first time I appeared on that who-is-who list was a year ago, and ZDNet reported on that fact under the humorous headline “Anti-patent campaigner hailed as IP hero”. While I’m not anti-patent in all respects (only against software patents), …

In my previous blog entry I described some configuration tasks that make the pre-configured pentaho demo environment installation act more like a server.

In this entry I will describe how to add a folder for our own MySQL samples. We will use this folder to deploy our real samples to. We want to use our own folder for two reasons:

• 
  
• presentation - it's much nicer if our samples stand out among the common ones everybody knows already
  
• maintenance - we can easily remove, rename or do whatever we want with all of our samples without affecting the other ones
  

Let's start off right away. Oh yeah, we won't need to stop pentaho - in fact, to demonstrate that, make sure it is started right now. Remember, you can start the pentaho BI …

MySQL on GNU/Linux appears to be able to either run multiple processes, or one process and multiple threads. We’ve noticed a significant CPU penalty for multiple processes, probably from the context switching overhead. The trouble was, one of our servers wouldn’t use threads; it wanted to use multiple processes. This article explains how we got it to use threads instead. First, we noticed the master server’s CPU utilization was higher than the replica’s, even though we expected the type of queries running on the replica should have caused it to have around the same CPU utilization as the master.

Probably the most frequently asked question to the docs team at MySQL from the public is ?I want to translate the manual into [insert language]?. That language can be anything from one we already have, through to some comparatively obscure suggestions.

read more