rra at stanford.edu
Thu Oct 19 22:23:37 UTC 2000
Nicholas Geovanis <nickgeo at merle.acns.nwu.edu> writes:
> On 19 Oct 2000, Russ Allbery wrote:
>> I wish I knew why people were seeing these anomalously high history
>> write times, as I'm not seeing this on either of my machines. My test
>> server (which is a dual-processor SPARC 20 and therefore is currently
>> running pretty close to flat out) sees:
> Maybe a dumb question, but is this an IDE vs. SCSI issue?
Hm. That's a good question.
For comparison purposes, headwall (the first one listed) is writing
directly to disks in a Sun SSA connected via fibre; it's not using RAID of
any sort, nor is it using SCSI.
The second machine, newsfeed, is using a small bit of an EMC Symmetrix and
hence is writing mainly to a 4GB writeback cache across UltraSCSI.
Russ Allbery (rra at stanford.edu) <http://www.eyrie.org/~eagle/>
More information about the inn-workers