Saiweb

Tweet

So I thought maybe it’s time for an update.

The project is moving, albeit slowly, and I realy do not like the current PHP implementation, I want to move towards a C++ version, and I will do so as soon as I figure out how to do CLI “update/refresh” …

i.e. See how W GET works with the progress bar and kbps all in text, I have NO idea how that works.

I am also looking at adding RRDTOOL support.

Anyway here’s the current sample output:

Detail:

… (I’ve truncated this very long list)

… (Truncated again)

.. (truncated)

Tweet

Seems that dbStat had to be developed much faster than I realised, it _was_ a personal project that I was working on as I got time.

Now it has been completely re-written to provide a complete “break down” of a large mySQL database suffering some major iowait (on average 15%).

The database in question was 17GB in size with 63% of that data size being pure indexes … MAJOR headache.

After completion of v1.0 dbStat, and subsequent review of the output, we were able to reduce the Index size by 7GB (41% of Total).

So the datbase is siting at ~10gb (40% index), plenty more work to do, but by removing the problem causing indexes we have reduced the index overhead and stemmed a growth of ~380mb/24hrs

Rapid development has left a few bugs, currently dbStat is at v1.2 with a very buggy CSV export function, anyway more as the project progresses, it also raises a few questions about public release now due to it being laregly to resolve a business issue.

Hmm maybe I can get a testimonial! haha

EDIT: iowait is down to 2.28% now

Tweet

I have been working on a script as of late to aid in the ever ongoing process of optimizing a web applications “back end”, inevitably the database, and underlying code.

Thus dbstat was born, the current version is not for release _just yet_ , as some tweaks are still needed, at the moment it only supports reporting for a single database at any one time.

v 0.5 Features list.

1. Dynamic / Fixed row size checking. (Dynamic size rows can cause table fragmentation).
2. MyISAM / InnoDB  checking (Reports total number of tables using these engines).
3. Index threshold testing.
4. Table fragmentation threshold testing.
5. Table size threshold testing.
6. Database size reporting.

At the moment this is currently written in PHP, and is quite heavily scripted I will be looking to move to C++ for a v1.0 for the release, for calculations sake and to remove the need to install PHP.

I will post more information as the project develops, I will be looking to release the program and source under GPL by version 1.

By version 1, I intend to have included the following:

1. User interactive reporting. (On even of threshold being exceeded, prompt user to fix or skip, with cursory notes*)
2. Table de-fragmentation.
3. Reporting to include suggested fixes.

*Cursory notes  to describe the fix, and possible implications of actioning it.

So watch this space, and please leave requests via comments Tweet