Developers of Sympa

[Olivier Salaun <address@concealed>]

Fil wrote:

> I really don't understand why it takes up to 5 minutes just to "find" one
> address in a list that contains 120.000 addresses.
> 
> If I check the database directly with the mysql client, it takes 3 seconds
> or less to ask something like
> 
> select * from subscriber_table where 1 and subscriber_table.list_subscriber
> like "info-diplo" and subscriber_table.user_subscriber regexp
> "florence" LIMIT 0,30;
> 
> Unfortunately I am unable to get mysqld to log the requests, so I cannot
> check what's going on.

It's a shame because it would help understand the problem. Did you try
setting
--log=/var/log/mysql.log option at mysqld startup ?

Currently I see 2 reasons for such bad perfs :

        1/ Sympa does some queries twice
        => we have added a in-memory cache in CVS version

        2/ Sympa does joins between user_table and subscriber_table
        => we have changed this in CVS version


** Note that CVS version should only be used for development/test needs
**
** CVS = unstable                                                       **

--
Olivier Salaün
Comité Réseaux des Universités