Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 

   


   


   















      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] Re: Re:Computer size: RAID



[This is a bit off-topic / harddisk-technical]

I wrote:
>> [...] HPT RAID-chipsets [+] Seagate Baracuda IV [...] *lousy*
throughput


On Saturday, May 11, 2002 10:02 PM, "Austin Franklin"
<darkroom@ix.netcom.com> wrote:

>If you have this system, have you corresponded with HPT or the
>manufacturer about this issue, or talked about it in a newsgroup?

Only discussions on StorageReview, where my nickname is pesky.
http://forums.storagereview.net/viewtopic.php?t=430

Having wanted for some time to play around with the RAID chip on my
motherboard, I bought two drives at the local small Aldi (grocery) store
:-)

Then I ran a few RAID-0 tests, and found bad performance. Then I
discovered (via StorageReview) that I was not alone.

The popular (=cheap) on-board HPT controller chips depend on the drives
to buffer data, but that's IMHO okay. I ran a lot of boring low-level
tests, trying to guess what was wrong. It looks as if the 'Cuda IV
flushes it's buffer if it isn't emptied as fast as the data is read
sequentially from the platters. With two 'Cuda IV drives we are talking
110-120MB/sec, i.e. close to the maximum burst throughput of a 33MHz PCI
bus. The combination of a non-caching RAID controller and a pessimistic
disk buffer algoritm results in lousy performance - the drive ends up
having to rotate the spindle multiple times to get a single track read
into computer memory. Not good.

It would work well with a 66MHz PCI bus, or with a caching RAID
controller... but Seagate's press release still says (no errata notice)
that the 'Cuda IV is a good choice for entry-level RAID:
http://www.seagate.com/cda/newsinfo/newsroom/releases/article/0,,1156,00
.html

> It could very easily be simply a set-up issue?

Yes, but it isn't, it's a well-known problem. I now see that Seagate
offers to replace drives with a special version - but only people who
*complain* are actually told about this:
http://forums.storagereview.net/viewtopic.php?t=2464
The drives in stores are still lousy for RAID-0, and buyers are not told
about the bad RAID-0 performance - owners will have to discover and
identify the problem themselves...

I am now using one of the drives (standalone) in my Linux server, so I
can't just RMA it. Furthermore, I have mostly given up complaining
through foreign manufacturers' first-level defence line AKA "customer
support". They tend to stress my mental sanity.

/Peder Skyt

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body



 




Copyright © Lexa Software, 1996-2009.