> Do you use UDFs? Faulty UDFs can cause things like this, malfunctioning hardware or crashes when running without forced writes are other possibilities (I think)

in the past (3 week ago) yes a lot ... we finally remove all of them by the internal firebird function or stored procedure... but as you see it's not answer the problem :(

> >server freeze
> Poor transaction handling can make things come to a crawl. What are the database statistics when you observe this (when things are slow, not after restarting), of particular interest OLDEST (ACTIVE) TRANSACTION and NEXT TRANSACTION.

All the transaction are like this

procedure execute sql
start transaction
do the job
commit transaction
except => rollback transaction

so transaction are just open for the time to execute the query and close as soon the query is over :(

also the database connection are in a pool, and are close if they are not used for more than 20 min

> Well, our databases ran fine when our database administrator got a disease a few years back. I might have restarted the server every now and then (though not every month) and done a few ad hoc backups and restores, but Firebird ran virtually without any administration. Though I admit that our databases are small (the main tables containing between 1 and 15 million records) and normally only used by 20-30 persons at a time, so I'm not amongst those that have experience with Firebird on larger systems (although I haven't heard Sean Leyne complain, and I think his systems are pretty big).

i don't know what to say execpt : "Why me" :( i m a good guy why firebird not like me :(