Re: disk layout + performance

From: Ong Beng Hui <[email protected]>
Date: Tue, 11 Feb 1997 01:05:43 +0800 (SST)

> I'm not sure I agree with your "tieing all the heads together".
> (Maybe that's some oddity of the OSs you're familiar with?). With

Typical implementation of RAID-0. Perform a spindle lock
on the all the disks to be stripped and basically tie
the heads together. Correct me I am wrong. I think that's
how RAID-0 works. If the controller has a large write
back cache, it might offset the effect.

Squid access is very random, a large percentage
of time for disk access will be spent on seeks instead
of read/write data. Since, the head are already tied
together when you stripped it, independent access
to different location of disks can't be performed.
Received on Mon Feb 10 1997 - 09:17:13 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:34:25 MST