Milan Babuskov wrote:

> So, maybe a gun with 1000 cylinders could be a better comparison - if
> there weren't for 'ordered writes'. If you use a filesystem with ordered
> writes (ext3, reiserfs since kernel 2.6.9, etc.), chances of corrupt
> database are really low. You can lose some committed transactions, but
> you won't get corrupt db file.

Do you mean that some file systems (on Linux?) have a write cache but
ensure that writes are performed in the order they are queued?

Or are you referring to a FB feature?

In either case, I assume the disk still has to perform the same sequence
of seeks and writes to do its job, so with a constant commit rate the
performance will be the same as without this kind of cache. Same
"motorcycle effect". Right?

Kjell
--
--------------------------------------
Kjell Rilbe
DataDIA AB
E-post: kjell@...
Telefon: 08-761 06 55
Mobil: 0733-44 24 64