On Wed, 02 Sep 1998 12:25:09 -0700, Hans Reiser <reiser@idiom.com>
said:
> I see it as important that the API not perform the aggregation of
> each piece of resource data into a single file for the case of
> ReiserFS. Do we have consensus on this? Stephen?
Nearly. I see it as important that the filesystem not care about
whether it's one file or several, it should still be fast. ReiserFS
can already deal with this without any new semantics. In the case of,
say, the KDE "parts" framework, they are already using MIME: we don't
want to force a new API on them. If they decide to supplement that
with GUI tags, then let them use subdirs, sure. It is not the job of
the fs to decide that API; it is the fs's job to provide a sufficiently
general and efficient interface upon which the application can build
whatever storage structure it wants.
> I don't know anything about RMS, but ORACLE uses raw disk for a
> reason..... To avoid letting layering give you the worst performance
> aspects of two storage models combined.
Actually, having spoken with them recently, it turns out that Oracle
don't really care too much about raw devices for performance (and I
heard somebody mention that they in fact recommend that you do not use
raw access at all in some configurations). What they really want raw
devices for is for parallel server, where they have shared scsi
devices and where they need to be very sure that they have got
unbuffered access to the device.
--Stephen
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html