Has anyone done any more work on recompiling InnoDB with 4k pages and benchmarking under SSD?

We’re building out a new DB that uses very small records (around 32-64 bytes) so reading a whole 16k for this record should have a performance difference.

I haven’t seen any benchmarks on 16k random read IOPS on the Intel SSD but my hunch is that there will be a 20-30% penalty here.

Though even if it was a 4x penalty that would still be about 9k transactions per second which is pretty good.

On a personal note I just bought a new Mac Book Pro which will be upgraded to the Intel X-25M MLC SSD.

Needless to say I’m very excited!

…

I deployed a mysqld binary with support for show table_statistics and quickly realized I need to improve it by replacing the column Rows_changed with the columns Rows_updated, Rows_deleted and Rows_inserted. I also need to support this as a table in the information_schema.

With the current code, I was able to find a performance problem. I setup a server as a slave that did nothing but replication and then compared the list of the top-10 tables by Rows_read and Rows_changed. One table had 55% of the rows read but only 1% of the rows changed. From there I found the SQL for that table in the relay log and found the inefficient SQL. Hopefully we can fix it.

The problem was an update statement that queried many more rows than it updated. I wonder if there is an easier way …

It has been more than a year since my self-imposed hiatus from serious MySQL development started and I think it is about time that I get back into the saddle. I have a handful of working prototypes but I should get the code out there, back into the community.I learned a bunch of stuff during the past year at Google but in the end, working on JavaScript, HTML/CSS and Google proprietary languages

A couple of days ago, Baron Schwartz posted some simple load and select benchmarking of MyISAM, Infobright and MonetDB, which Vadim Tkachenko followed up with a more realistic dataset and interesting figures where MonetDB beat Infobright in most queries.

Used to the parallel IEE loader, I was surprised by the apparent slow loading speed of Baron's benchmark and decided to try and replicate it. I installed Infobright 3.2 on my laptop (see, this is very unscientific) and wrote a simple perl script to generate and load an arbitrarily large data set resembling Baron's description. I'm not going to post my exact numbers, because this installation is severely …

Over the past few years I have found that "pager" inside of MySQL is a really useful tool. I have come up with a few simple, but extremely effective, ways to use it.

-- Example 1
Lets say you have a MySQL server that is really busy with extremely long queries. You run a “show full processlist” and everything going on streams before your eyes, new lines and all. A more readable way to see what is going on is…

Mysql> pager less –S

By running the above statement the full processlist will NOT appear in the traditional, word wrapped format. The output will show in a very readable, right extended, format. All you need to do now figure out what is going on.

Here is a helpful hint. Set a tee file for MySQL client output.

Mysql> \\\\T /location/to/file.txt

Then, make sure pager is set to “less –S”. From there you can get out of the …

Accidentally me and Baron played with InfoBright (see http://www.mysqlperformanceblog.com/2009/09/29/quick-comparison-of-myisam-infobright-and-monetdb/) this week. And following Baron's example I also run the same load against MonetDB. Reading comments to Baron's post I tied to load the same data to LucidDB, but I was not successful in this.

I tried to analyze a bigger dataset and I took public available data
http://www.transtats.bts.gov/DL_SelectFields.asp?Table_ID=236&DB_Short_Name=On-Time about USA domestic flights with information about flight length and delays.

The data is available from 1988 to 2009 in chunks per month, so I downloaded 252 files (for 1988-2008 years) with …

Does a query with FORCE INDEX reduce the overhead from query optimization? For me it did not when using sysbench. I configured sysbench to only do primary key lookup queries and ran it with the client and mysqld on the same server. Query optimization overhead is significant for a memcache-like workload (primary key lookups on cached data). I will publish more data on that in another post.

For QPS versus concurrent sessions the results are here

I modified sysbench to use a force index(PRIMARY) hint in …

This entry is a follow up to TOTD #95 and shows how to use the recent integrations of JSR 299 in GlassFish v3 to convert a JSF managed bean to a JSR 299 bean (aka Web Beans). The TOTD #95 describes a simple Java EE 6 web application that uses Java Server Faces 2.0 components for displaying the results of a database query conducted by EJB 3.1 and JPA 2.0 classes.

The EJB class, which …

We’re looking to hire a Senior MySQL DBA over at Spinn3r.

You should obviously have MySQL experience. Love SQL, hate data corruption and slow queries, and preferably live in San Francisco.

Linux experience would be nice as well but not required.

Extra points if you are excited about SSD, *huge* amounts of data, have hacked on Drizzle or XtraDB

Spinn3r is a GREAT place to work. We’re growing fast and have cool new offices in SOMA.

…

Sometimes we have a lots of small and rather fast queries which use group by/order by, thus creating temporary tables. Some of those queries are retrieving text fields and mysql have to use disk (myisam) temporary tables. Those queries usually run for less than 1-2 seconds, so they did not get into slow query log, however, they sometimes add serious load on the system.

Here is the stat example:

bash-3.00$  /usr/local/mysql/bin/mysqladmin -uroot -p -i 2 -r extended-status|grep tmp_disk
...
| Created_tmp_disk_tables           | 109           |
| Created_tmp_disk_tables           | 101           |
| Created_tmp_disk_tables           | 122           |
...

40-50 tmp_disk_tables created per second

So, how can we grab those queries? Usually we have to temporary enable general log, filter out queries with “group by/order by” and profile them all. On solaris/mac we can use dtrace instead.

Here is the simple script, …