Date: Mon, 19 Apr 1999 14:19:29 -0700 (PDT) From: Matthew Dillon <dillon@apollo.backplane.com> To: Jonathan Lemon <jlemon@americantv.com> Cc: Arjan.deVet@adv.iae.nl, hackers@FreeBSD.ORG Subject: Re: Directories not VMIO cached at all! Message-ID: <199904192119.OAA90535@apollo.backplane.com> References: <local.mail.freebsd-hackers/199904171844.LAA75452@apollo.backplane.com> <local.mail.freebsd-hackers/199904191650.JAA24137@vashon.polstra.com> <199904192058.PAA16517@free.pcs>
next in thread | previous in thread | raw e-mail | index | archive | help
:Which is why Peregrine prefers to use raw disk partitions (along with
:a userland variant of LFS) to store the pages, since the filesystem
:currently imposes too much overhead for good performance.
:
:It's interesting, LFS seems to be a great web-cache filesystem, you
:don't really need to preserve every file, you just throw some away.
:No fsck; if the system crashes, you can just start all over again;
:after all, it _is_ a cache, right? (In Peregrine, this behavior is
:tunable; some environments don't want to lose the entire cache).
:--
:Jonathan
Going through the buffered block device for a partition would
be quite efficient. FreeBSD can parallelize block I/O very nicely -
there are no locks to get in the way. Using a file isn't terrible,
though. You just have to use more then one. I used this trick in
Diablo ( my usenet news transit system ). Each file caches multiple
articles. It doesn't take a very large ratio to achieve optimal
performance. By limiting the number of files, you virtually guarentee
namei cache hits.
-Matt
Matthew Dillon
<dillon@backplane.com>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199904192119.OAA90535>