Joerg Schilling schrieb am 2006-01-30: > Matthias Andree <[EMAIL PROTECTED]> wrote: > > > > > 2. scheduling writes is a complex matter in itself, and there is not a > > > > single answer. The later you start writing, the more data you have > > > > available if the input rate is low for a while to write a large blob of > > > > data (reducing seeks), the more opportunity you have to rearrange writes > > > > and the lower the total amount of data to write, because updates of > > > > yet-unwritten data do not turn into writes (I/O) but are made in-core. > > > > > > If you have the buffer cache at the right place and if you are able to > > > cluster to already pending requests (like on Solaris), things work > > > different. > > > > You said Linux would schedule writes after too long a delay, not that it > > were incapable to cluster. What's going on here? > > I encourage you to have a look at the Solaris sources if you have problems > to understand clustering.
I don't care. As I wrote previously (which you didn't quote): talk to the right people if you have something important to say about the buffer cache that might improve your application. -- Matthias Andree -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
