Re: RAID 5 performance

dunk_at_mincom.oz.au (Duncan Young) wrote:
>In <46eoe6$e8v_at_ixnews2.ix.netcom.com> rstrouss_at_ix.netcom.com writes:
>
>>Paul Baumgartel <paulb_at_pcnet.com> wrote:
 

>>>I've just done a direct comparison of RAID 5 and non-RAID disk configurations.
>>>All disks are mounted on the same CPU, a Digital Alpha running OpenVMS 6.1. The
>
><snip>
>
>>>YMMV, but this makes it pretty clear to me that, for this application at least, RAID
>>>5 is not the way to go. Next I'm going to test RAID 0+1 (striped and mirrored), and
>>>I'll post the results.
 

>>We're setting up a RAID 5 envirronment on an HP 9000 box and I'm
>>hoping the 32 meg cache will minimize the problem that you're
>>encountering in your situation.
>
>Raid 5 without caching in not a valid comparison, of course its abysmal. If
>you have caching and data coalesing on separate hardware, the difference, if
>any, should be negligable.

Paul, please repost your RAID 5 and RAID 1 results. We're very interested in your findings. For us, RAID>1 is imperative due to availability requirements. The issue for us is whether to use RAID1 or RAID5. Of course RAID 1 costs significantly more, but what are the performance differences? Will we take a hit using RAID1 versus RAID5? We don't care if comparisons are made with caching or not as long as both are tested the same way.

Thanks. Received on Thu Oct 26 1995 - 00:00:00 CET