As any serious MySQL DBA knows, MySQL can and will now and again go out of sync, most of the time it cures itself, but sometimes the worst happens and a slave stops replicating. Yet I bet most apps keep on going with random outcomes, especially if your storing the users sessions in MySQL and have a mastermaster setup.

To counter this problem I implemented a quick an easy check. If the check catches a replication failure `touch` a file to the nfs mounted tmp directory that is shared by all nodes.

Each node before executing anything (in index.php) checks for a file called 'halt' in tmp. If found exits and bingo no weirdness, corruption, or wondering why something is there one click, gone the next. It even emails you, and with a simple addition can sms.

mysql_query("show slave status;", CONNECTION);
if(mysql_errno(CONNECTION)!=0)
{
   mail('team@<some url>',
 …

I stumbled across this on the fabulous http://www.everythingsysadmin.com :

from:  http://www.socallinuxexpo.org/wios07/

The Southern California Linux Expo (SCALE) will host a Women in Open Source Event as part of their upcoming 2007 conference, SCALE 5x.

The focus of this event is on the women in the open source and free software communities. The goal of this event is to encourage women to use technology, open source and free software, and to explore the obstacles that women face in breaking into the technology industry. The Women in Open Source event will be held on February 9, 2007 at the Los Angeles Airport Westin Hotel.

I have seen the dearth of women, in system administration as …

A very interesting document from BBC's Standards and Guidelines (Database Design and Development Standards v1.00) mentions only two databases, guess what? Oracle and MySQL!!!
BTW the document is also a source of good design tips like:

Normally, use only alphanumeric characters plus the underscore (_) character for table names.

MySQL
Table names are case sensitive under UNIX (or Linux). To avoid confusion, define/declare table names in upper case in any database creation SQL scripts.

No SQL language reserved words should be used as column names

Identify primary, unique, and foreign keys (to enforce important or critical data integrity) as part of the initial design process

To guard against slow …

OK, so I stumbled upon Working With Rails a while back when there was a little controversy about their methods for populating the database with Rails developers and I figured I should claim my record. I filled in a few details then forgot about it. I visited again today to mention that I’m a Rails core contributor (well, I’ve had a patch or two accepted, which is good enough for me!) after reading that CD Baby is sponsoring conferences passes for the top 20 contributors between now and the conference registration opening for RailsConf ‘07. I sincerely doubt that, with my current workload, I’ll either contribute much or make it to the American RailsConf, but why not sign up anyway?

While I was visiting …

Seems like the MySQL Symposium will be rescheduled. On one side that's bad news, on the other personally it makes my schedule a bit lighter.

On the same day I got the news about the MySQL symposium being delayed I got the news that my paper for the dutch Virtualisation conference has been accepted at least I have more time now to work on that :)

An excellent opportunity has come up to have the 2007 edition of the Open Source Developers Conference to Brisbane. Various people from Brisbane were independently interested, and we converged at LCA 2007 in Sydney last week. From discussions there, I understand that the OSDC committee is interested, but would of course like to make sure that it will once again be a well attended quality event, and a worthy successor to the previous editions in Melbourne - so we need to do some homework!

OSDC is generally held in December, the CfP would start in April. Various things would need to be in place before then. If you are able and willing to help with this, please email me at my-first-name @ my-last-name dot com dot au The current aim is for an initial meeting in Brisbane within the next couple of weeks, to go over things before submitting a full proposal to the OSDC committee.
…

Some might've noticed that the check-cpu script in the MySQL BUILD directory does not always succeed in detecting the right processor:

roland@roland-laptop:/opt/mysql/5.1/bitkeeper/BUILD$ ./check-cpu
BUILD/check-cpu: Oops, could not find out what kind of cpu this machine is using.

When you dig a little deeper into that script, you will notice that the name of you particular processor (in my case: Intel(R) Core(TM)2 CPU T5600 @ 1.83GHz) is indeed not there and thus not checked. Finally you may find that the model and name of your processor are passed to the gcc compiler.

As far as I cans see the information is used to set the values of the --mtune and --march compiler options. What those are? Hehe! Ok, here's a partial result from man gcc:
…

At the end of my last MySQL post I mentioned strange behavior with GROUP BY and DISTINCT. In Oracle such constructions are not valid and produce an error. Here is the test table and statements:

create table t1 (c1 int, c2 int);

insert into t1 values (1,10); insert into t1 values (2,20); insert into t1 values (3,30); insert into t1 values [...]

I hear a lot of people say: "Well, compiling MySQL from source might pose a problem to Windows users, but it's straight forward for Linux users." Quite often, the difference is explained by claiming that Linux users are used to doing things themselves, and gaining an extra bit of customization by doing so.

My impression is that a large part of the people that say this, are already quite experienced Linux users that are quite comfortable with compiling software from source. The entire toolchain that is involved is already set up on their system, and some of them probably already forgot what they had to do in order for it to make it work.

What is quite often not mentioned is that the required toolchain does not magically land on your system. If you are experiencing prolems compiling MySQL source it's quite likely the toolchain …