Hello,
People's experience of using ccd on newsservers shows that
it's better to have larger interleave for such workload.
IMHO, Squid does not differ too much from the newsserver,
because the average object size is larger than the average
article size only by one order of magnitude (15-20K vs 2-4K,
not considering alt.binaries.* :-) and its disk activity
has a quite random manner. Manpage ccd(4) says:
For random-access oriented workloads, such as news servers, a larger in-
terleave factor (e.g., 65,536) is more desirable. Note that there isn't
much ccd can do to speed up applications that are seek-time limited.
Larger interleave factors will at least reduce the chance of having to
seek two disk-heads to read one directory or a file.
Peter Marelas wrote:
> Im using it on 2 systems running 2.2.2-RELEASE.
>
> It works great..my config is..
>
> ccd0 8 none /dev/sd0b /dev/sd1b
>
> The partitions are 4GB each.
>
> The load balancing works well and provides good performance.
>
> On Thu, 13 Nov 1997, Francis A. Vidal wrote:
>
> > hello everyone!
> >
> > i would like to know how many of you use FreeBSD's CCD striping with
> > Squid? can you please share some experiences? thanks!
> >
> > ---
> > u s l s N E T university of st. la salle, bacolod city, philippines
> > . . . . . . PGP key at ftp://ftp.usls.edu/pub/pgpkeys/francis.pgp
> > francis vidal tel. nos. (6334).435.2324 / 433.3526
> >
>
> Regards
> Peter Marelas
> --
> Phase One Interactive - Sun Solaris/Unix/Networking Consultant
> P.O Box 549, Templestowe 3106 Melbourne, Australia
> URL: http://www.phase-one.com.au/
>
SY, Yar
Received on Thu Nov 13 1997 - 01:11:20 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:37:31 MST