$ cat /etc/my.cnf

datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
user=mysql
# Default to using old password format for compatibility with mysql 3.x
# clients (those using the mysqlclient10 compatibility package).
old_passwords=1

max_connections         = 8000
wait_timeout            = 300

log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid

$

It’s going to be a long report . . .

I have long held the opinon that SSD (Solid State Disk) drives are going to be a major part of the database future. I just checked and I wrote a blog posting about them two years ago. I am not alone in this opinion.  It has long been realized that both I/O access speed and throughput increases have not kept pace with the increases in CPU power and the steadily decreasing cost of RAM. Storage space has increased, but both access speed and throughput performance have only had marginal increases in performance.

Solid state disks have long held the promise of lowered access speeds, especially when it comes to random access.  Even so, prices for SSD drives have been high and space small (compared to standard hard disks). That is slowly beginning to change. I suspect that within five years SSD drives in RAID arrays for database servers will be just as common as standard disk arrays are today.

Pliant Technology announced today what they call an Enterprise grade series of SSD drives that they claim has twice the throughput of the leading competitors. Computer World has a good writeup of the announcement here. Interesting technology that Pliant claims can produce a maximum of 180,000 IOPS from a single 3.5? SSD drive.  While undoubtedly expensive, for those who are I/O bound with few other options this could be a God-send.

Be prepared . . the time is coming sooner than you think that SSD drives are your main system drives and traditional hard drives are used for backup.

A bit of googling uncovered Bitvise’s Tunnelier. It is a free for personal use tool that is designed just for tunneling. It seems to handle it with some aplomb. Ten minutes after I downloaded it, I was online with a remote server using MyQuery.

If you are using a Windows desktop and want to use some of the various graphical client programs that don’t include port forwarding this could very easily fill your needs. I just wish I had found it a long time ago.

Daniel Wall, lead legal counselor for Oracle, said in a prepared statement on June the 26th:  “we were almost able to resolve everything before the Second Request deadline. All that’s left is one narrow issue about the way rights to Java are licensed that is never going to get in the way of the deal.”

I guess he was wrong.

Oracle has said publicly it wanted to close the purchase of Sun by the end of August which is the close of Oracle’s financial quarter. Since I am posting this on September the 3rd it can be confidently said they didn’t achive their goal.

In late July 62% of the Sun stockholders approved the Oracle acquisition of Sun at a stockholders’ meeting. While it was certainly a majority, I suspect that the involved parties wished the percentages where a bit higher. It appears that everyone is not the Oracle bandwagon. Even so, on August the 20th the DOJ approved the merger.

The last thing Oracle wants with the Sun purchase  is a protracted legal battle similar to the PeopleSoft acquisition. This transition has already hurt Sun’s bottom line and talent is leaving Sun at an alarming pace. Of course the MySQL talent diaspora is well documented, but others are leaving as well. This includes the recently departed JRuby team who left to join Engine Yard. Despite what some companies might think, the most important part of any acquisition is the talent . .the people. I don’t have the time or space to turn this into a post on the employees, but I think it is a very valid concern.

This brings us to today. The only problem remaining is for the EU to approve the merger.  The announced delay is because the commission wants to investigate the issue of MySQL competing with Oracle (the database server). I thought it would be helpful to outline the possible outcome of the EU commission’s review. Results of the is review are required to be released within 90 days working days (January the 19th, 2010).

This verdict has three possible outcomes:

1) approval with no changes

2) conditional approval with required changes.

3) Outright rejection

Those changes could be anything ranging from requiring a change in the license terms to spinning off of MySQL and associated products into a seperate subsidiary or even outright selling of MySQL and it’s assets.

If there are any changes required for approval, Oracle will essentially have to give in to any requirements. As a large multi-national company who sells something like 60% of its products/services outside the United States it is going to have to cooperate with this commission in the long run. Oracle might choose to fight it legally if it can but I think that the end result will be that Oracle will be making the necessary changes.

If you are affected by this decision it might be wise to take action and let the EC  know what you think. If you have any concerns you should contact Neelie Kroes’ office in the EC.

$ cat /etc/my.cnf

datadir=/var/lib/mysql
socket=/var/lib/mysql/mysql.sock
user=mysql

log_slow_queries=/var/lib/mysql/slow-queries.log

# Default to using old password format for compatibility with mysql 3.x
# clients (those using the mysqlclient10 compatibility package).
old_passwords=1

log-error=/var/log/mysqld.log
pid-file=/var/run/mysqld/mysqld.pid

max_connections = 600

# myisam options
key_buffer_size = 16M
myisam_sort_buffer_size = 16M
join_buffer_size = 1M
read_buffer_size = 1M
sort_buffer_size = 2M

#innodb options
innodb_buffer_pool_size = 5G
innodb_file_per_table
innodb_log_file_size = 128M

# miscellaneous
log-bin
expire_logs_days = 7
table_cache = 512
thread_cache_size = 100
wait_timeout = 1800
connect_timeout = 30
max_allowed_packet = 16M
max_connect_errors = 1024

# defaults
default_storage_engine = innodb

Server is up and running but doesn’t show the right values for some variables. What’s wrong? I had to scratch my head for a couple of minutes on this one as it wasn’t immediately obvious (to me). Nothing was logged in the error log…

Answer in comments!!

Oh dear. I was hoping that by now this would be a non-issue. It seems that many of us (including the majority of my clients) are late to the party.

People…it’s now September of  2009. Next month starts the last quarter of 2009.  If you have a 64-bit chip in your server for the LOVE OF GOD please run a 64-bit operating system. And easily 90% of the servers shipping have 64-bit procs. I don’t know of any shipping dedicated servers that are running 32-bit procs but I suppose there are some left somewhere in the universe.

Do you know how difficult and costly it can be to upgrade your server to a 64-bit OS? Costly in terms of money and downtime (which may as well be money). All because you didn’t realize when you bought the server that you would ever use more than four gigs of RAM. Or you thought you might have some trouble getting your application to work with the 64-bit libraries or whatever. Data sets can increase at almost exponentially rates and your server performance is almost a direct corollary to the amount of RAM you have available to utilize for the database server. What worked for you even six months ago might not be enough for you today. And without jumping through some serious hoops there is no effective way to use memory above 4-gigs on a 32-bit system.

The new Dell R710 (which is replacing the 1950) has a baseline price of around $1500. This server will support  two quad core Xeons and 144GB of RAM. Yes, that is correct, 144GB of RAM. It has 18 slots for memory. This will be NORMAL SERVER in months..not years.

Please raise your right hand and solemnly swear  ”I will never put anything on my database server but a 64-bit operating system.” Thank you for your cooperation and congratulations you just saved yourself both time and money. Now if someone could just order me a dell R900 please…

</rant>

The feed for the podcast is at osdbzine.net/podcast/osdbpodcast.xml and a direct download of the episode is available here.
Episode Summary
This is the first episode of Open Source Database Podcast. This pod cast is about both open source database servers and the people who develop and use them. The plan is for most episodes to be centered around interviews with various people in the open source community. Kicking things off with this inaugural episode is an interview with Brian Aker who is the creator of the Drizzle database.
I will apologize now for the beginning and ending of the podcast where the recording level is too loud. My only excuse is I have new equipment I am working with and I am not used to it yet!

Resources
Sun’s announcement about purchase by Oracle: sun.com/aboutsun/pr/2009-04/sunflash.20090420.1.xml

MySQL Server 5.4 overview: dev.mysql.com/tech-resources/articles/mysql-54.html

Postgres Website: postgresql.org/
Brian Aker’s blog: krow.livejournal.com/
Drizzle: drizzle.org
Drizzle code: launchpad.net/drizzle
Open Source Database Magazine: osdbzine.net

Production Notes
Recorded using Audacity and an Alesis MultiMix 8 USB mixer. The interview was recorder over Skype with the Alesis mixer.
Release date: July 24th, 2009

So it takes 10 (business) days to update your database. You really might want to investigate MySQL. It works overnight and on the weekends. And it sounds like it would be quite a bit faster!!

A great article on the new features of Postgresql 8.4 by Robert Treat
Part one of a two part article on Percona’s new XtraBackup backup program

Also the news, the book shelf and Peter Brawley’s Coding Corner. Downloads are available at osdbzine.net.