Hi all,

I wrote a real time application (in C++) which uses extensively Mysql DB. We have a specific ext3 partition for mysql. Partition is big enough but we want to have a full control and avoid any possible crash due to out of space for DB.
How can I manage such critical event? Application runs in critical environments and it is crucial it continues running without breaks. It cannot crash and stop because of partition full.
I could catch partition full event, but when it happens it's too late to rescue my pending data to be saved in DB.

I was even thinking to use a temporary partition which will be automatically mounted and used ONLY when the main partition for mysql gets full (anyway don't know how to do it). All this process must be transparent to the end application and doesn't have to stop it. Then I can have a background check to recognize when temp partition is used. Upon such event then application will take specific chooses and manage all situation, eventually removing some old data from DB and turn back using the only first original partition.

Anybody has a better solution ?

Thank you
Luca