Hello
Le Vendredi 3 Décembre 2004 16:00, Plamen Gradinarov a écrit :
I'd like to, especially in the face of the fact
that after 15 hours of
intensive indexing, the disk with all server databases collapsed (there was
nothing else on it, only the mysql db files - about 15 GB used out of 120
GB).
You probably have _hudge_ logs for apache and mysql, in /var/log/
It logs each connection, and each request, which produce logs
bigger than database, and i think is ressource consumming.
here it takes more or less 2-3 hours, on Athlon 2200 + 750 MB
and a normal disk (40 GB, maxspeed 20 MB /s) for the
french text database (270 MB uncompressed, 60MB compressed)
and the index database is roughly the same size as the data database
(it is the same for htdig, and i think for every search engine, which put
all the data in its internal searchdatabase)
Here are my.cnf settings
I would appreciate any corrections and suggestions.
I have the same question , and need for good informations on
settings for the servers:
[mysqld]
datadir=/disk2/mydbs/mysql
socket=/var/lib/mysql/mysql.sock
skip-locking
skip-innodb
Myabe, you should comment the 2 above lines.
This seems to be Debian Specific. They said to read the doc
(mysql-doc /non-free) if you want to enable InnoDB (i ll read this night ;)
It seems, that it install as MyISAM (i d'ont know the difference,
but it is different), and you can change this using phpmyadmin.
It take a long time (one hour or so on athlon 2 GHz + 750 Mo)
and after i dont know where the data are !!!! but it seems
to be very much faster (12s for the special page "allpages",
instead of 5 min)
my 2 cents
alain