SCSI/Controller..Choices

[sdgserv]

Building a sys. Customer wants boot drive to be SCSI.

My choice
Fujitsu..Max 3073 WC/NP/FC/RC 73.5G
Lsi 3442 X-R

Asus P5 DH Delux wifi.. Mobo
My question is..
Good..Bad. Just OK? Talk him into a Raptor/

I am not a big fan for SCSI for Home use on PC. But I do what I am told.

Any suggestion for other SCSI drives or controllers..I am all ears

Thanks in advance

 

[defakto]

I'd try to talk him into a raptor if possible as it's aimed towards desktop usage and the scsi is more likely towards server usage patterns but, if they still want it, the customer is always right in their own mind.

 

[UICompE02]

Well, exactly, I'd have to question using SCSI on a desktop, though I do it myself (it helps that I get the controllers for free...)

But in this case, going for SAS, as you appear to have, is a wise choice since the customer can then add SATA drives for bulk storage with tons of expandability options. However, after adding just a couple drives to that controller, you'll very quickly run into the bandwidth limitations of the PCI 32/33 slots on that motherboard.

What are the customer's intended expansion plans for this system? Maybe it would be best to get at PCI-E SAS controller such as the LSISAS3442E-R and a motherboard with an extra x8 or wider PCI-E slot if the customer intends to expand the storage subsystem beyond a few drives.

 

[sdgserv]

Excellent question..I will ask him

Thanks

 

[sdgserv]

This guy is stubborn..whoa
anyway he is going to use 2 drives and now he is talking about striping.
I know I will back on this forum

 

[drizzt81]

Here's a suggestion for you:

"Dear Customer X,

I understand that you would like me to build you a desktop performance gaming uber-1337 rig with a pair of SCSI HDDs in a striped set. While the sheer amount of money you will spend on this will certainly insure that you have the biggest ePenor on the block, I would suggest to take into consideration that there is a high probability that using a single WD Raptor will actually be faster for your workload, i.e. uber-1337 gaming, and cheaper at the same time.

I would suggest going this route, but if you insist on going the SCSI route, I am more than happy to help you in that area as well. My research in the SCSI scene has brought me to the conclusion that I need to ask you the following questions before we can proceed with it....

Your excellent PC supplier"

well, maybe you don't want to copy my thing ad-verbatim, but you get the idea. If you do this and he comes back in 3 months telling you how much his 1337 SCSI rig sucks for loading BF2, you can show him: ``I told you that we could have implemented a better alternative for cheaper back in the day. However, I am sure that in order to satisfy you, we can come up with an alternate solution: I will take your PC back, sell of the SCSI drives and controller and with the resultant cash, we will purchase an IDE setup for you.''

 

[sdgserv]

LOL..

I don't know you, but I like every thing you said. HA!

I have been building comps. on the side for some time now. And I sure don't know everything. I have just never had to deal with an ego like this one.

It only a computer.........

 

[drizzt81]

LOL..

I don't know you, but I like every thing you said. HA!

HI,

I am Greg, the local cynic. You can join me and the other fellows in the Off-topic area for a small fee. I appreciate the positive feedback. While I do not build computers for others, I can emphasize with your situation, given that we have plenty of arguments in this subforum when it comes to ``real-world'' HDD performance, particularly w.r.t. gaming. In the absence of a repeatable, real-world benchmark that does not require tons of user interaction, I have come to rely on SR's benchmark database. They claim to have real-world benchmarks, but their results cannot be reproduced, since the software they use does not exists (cannot be bought) anymore.

Given the above mentioned problems to get reproducible real-world numbers that gamers are interested in, the debate is doomed to go on for eternity; or until someone steps up and spends two to four weeks in a lab with a bunch of HDDs and a stopwatch, staring at CS:S, BF2(142) etc. loading...
Of course at that point people are then going to go on and doubt his numbers and unless we get someone else to repeat and verify these results....

Anyway, welcome to the Disk Storage subforum. I strongly suggest to stay away from any thread with the word "RAID" in it

 

[unhappy_mage]

They claim to have real-world benchmarks, but their results cannot be reproduced, since the software they use does not exists (cannot be bought) anymore.

"does not exist", Grammar Nazi And you can get a demo version from SR. It's a pain in the butt, but it does work.