Taskkill will be ok if no one user worked with yours database and no internal process worked with your database. And it's not OK when you have more than one database on such server, because you interrupt working sessions to another databases.

I corrupted databases very well with such method. It's like sudden reboot when users work.
Just try to write a lot of data to db and execute taskkill(maybe when you use very short transaction all will be ok, but if you have long transactions and a lot of data it will be problems). The result will be orphan pages or orphan records if all was good, but not any time all is good. And taskkill during autosweep/sweep is really dangerous. And after experiment check your database by gfix.

 

> If you execute taskkill /IM fb_inet_server.exe /F then you are kill your
> database(broken indexes, missing data or some else things. if autosweep is
> on, you may completely kill your database).

Not true at all!

We use taskkill all of the time to force a shutdown of FB instances when we need to ensure that all connection processes are cleaned-up -- for our 100+ client production systems.

You will never experience a corrupt database.


Sean

-- 
Regards,
Markov Dmitri