Re: Freeradius Mysql Performance
What?? You dont need that kind of hardware for job, sure. Throwing that kind of horsepower might fix the speed but this is a DBA question. Look at your mysql configuration and see how it can be adjusted (my.cnf) look at the engine in use and see if you can use better..(eg innodb instead of myisam), look at an alternative SQL eg postgres. Look at your usage of sql with freeradius, eg the radius tables. What indexes are present what do you need , what do you not need? Can you divide the work? Use one server for one table or task and the other another...eg simple queries can be done against a passive slave server... alan
On Sat, Jan 28, 2012 at 3:03 PM, Alan Buxey <A.L.M.Buxey@lboro.ac.uk> wrote:
What?? You dont need that kind of hardware for job, sure. Throwing that kind of horsepower might fix the speed but this is a DBA question.
Look at your mysql configuration and see how it can be adjusted (my.cnf) look at the engine in use and see if you can use better..(eg innodb instead of myisam), look at an alternative SQL eg postgres. Look at your usage of sql with freeradius, eg the radius tables. What indexes are present what do you need , what do you not need? Can you divide the work? Use one server for one table or task and the other another...eg simple queries can be done against a passive slave server...
alan
Hi, Sorry to pick into this with a short question. Just wondering, do you see performance increase using postgres instead of mysql? I would rather think the opposite, but must admit that I'm no db expert and have not much experience with postgres. Kind regards, Yves
YvesDM wrote:
Just wondering, do you see performance increase using postgres instead of mysql?
Yes. MySQL can be higher performance than older versions of PostGreSQL, if you don't do database writes. Newer versions of Postgres have similar performance to MySQL, with the benefit of allowing writes. i.e. the "MyISAM" driver is fast but unsafe. The "InnoDB" is slower but safe. Postgres has the best of both.
I would rather think the opposite, but must admit that I'm no db expert and have not much experience with postgres.
The main reason to use MySQL is familiarity. That, and MySQL cluster. For most normal systems, Postgresql is a better choice. Alan DeKok.
On Sun, Jan 29, 2012 at 11:36 AM, Alan DeKok <aland@deployingradius.com>wrote:
YvesDM wrote:
Just wondering, do you see performance increase using postgres instead of mysql?
Yes.
MySQL can be higher performance than older versions of PostGreSQL, if you don't do database writes. Newer versions of Postgres have similar performance to MySQL, with the benefit of allowing writes.
i.e. the "MyISAM" driver is fast but unsafe. The "InnoDB" is slower but safe. Postgres has the best of both.
I would rather think the opposite, but must admit that I'm no db expert and have not much experience with postgres.
The main reason to use MySQL is familiarity. That, and MySQL cluster.
For most normal systems, Postgresql is a better choice.
Alan DeKok.
Ok Alan, I will not immediatelly will change the whole thing (indeed familiarity and we have no issues with our tuned mysql so far), but I will sure keep this post in mind. Thx for the clear up. Yves
Hi,
Sorry to pick into this with a short question. Just wondering, do you see performance increase using postgres instead of mysql?
yes. I am a postgreSQL convert. though, that said - out of the box you get slightly better and safer performance - but you'll still have to configure things (eg indexes) properly...and update your skillset as postgres does some things differently. alan
participants (3)
-
Alan Buxey -
Alan DeKok -
YvesDM