HD Configuration Advice?

[RKBA]

Hello all,

I don't really have any specific questions, but am just looking for general recommendations I guess.

When I bought my system about a year ago, it had two 30 GB hard drives plugged into the Promise FastTrack100 controller running in Striped RAID0 mode, and had 256 MB RAM.

About a month ago, I upgraded the RAM and very fortunately it turns out, added a 60 GB hard drive to the primary ATA100 channel to use as backup for the two 30 GB striped set of HD's on the Promise controller. Last week, one of the 30 GB drives gave up the ghost, and I reconfigured things as shown below. I now have WindowsME on the 60 GB "C" drive (and am planning on "upgrading" to Windows 98 http://www.PCGuide.com/ubb/wink.gif ), and Windows 2000 on the 30 GB "D" drive. Fortunately, I was able to recover everything off the 60 GB drive with no ill effects other than loss of my Linux partition which wasn't backed up. http://www.PCGuide.com/ubb/frown.gif

I realize having both hard drives on the same ATA channel isn't the best way to go, but I couldn't bear the thought of plugging one of them into the EIDE channel - and besides that, because of the way the cabling and CDROM & ZIP drives are arranged inside the drive bays, it would be almost physically impossible to connect the 30 GB hard drive up to the EIDE channel unless I totally rearrange all the other drives or get some of the round EIDE cables that aren't as wide as the ribbon cables (which I plan on doing).

So anyway's, the obvious thing to do is buy a 30 GB drive identical to the one I already have (if they're still available), and go back to the same configuration I had previously. One better alternative I think, is to buy another 60 GB drive identical to the one I just purchased last month, and attach those to the Promise RAID0 controller. The 30 GB drive is a year old and may very well go out like the other one did. I was really surprised that it went out as soon as it did, because most hard drives I've had have lasted several years. It was a true hard failure though, because it started making a kind of "whining" noise as though one of the r/w heads were dragging on a platter.

One other thing that bothers me, is that while I still had the two 30 GB drives on the RAID controller operating in striped mode and after I had installed the 60 GB drive on the ATA/100 channel, I ran a Norton Utilities drive benchmark and it indicated that the single ATA drive was faster than the striped RAID set!!! I assume Norton Utilities was lying to me as it sometimes does, and wonder if anyone can recommend a good hard drive benchmark test. Oh, and the Promise RAID card does (and did) have the 80 conductor ATA cables on it that connect to the striped RAID drives, so they should have been running at full speed.

So that's the situation. Any thoughts or recommendations?

-- Ron

_______________________________
Mainboard: Asus A7V Socket A, w/ACPI Bios Rev 1004C, and VIA KT133 chipset.

CPU: AMD Athlon T-Bird 1.1 GHz

RAM: 768 MB SDRAM (Couldn't resist the cheap RAM prices these days ;-)

EIDE 1
Master: Pioneer ATAPI DVD/CDROM
Slave: Unused

EIDE 2
Master: Plextor CD/RW PW1210
Slave: Iomega Zip 250

ATA/100
Master: IBM 60 GB, IC35LO60 AVER07-0
Slave: IBM 30 GB, IBM-DTLA-307030

PROMISE FastTrack 100 Controller: Removed from system at present (see comments)

[This message has been edited by RKBA (edited 09-03-2001).]

[Paleo Pete]

Without doing any serious digging, the difference in speed might be in the transfer rate. The 60GB drive is the Deskstar 60GXP (http://www.storage.ibm.com/hdd/desk/ds60gxp.htm) which shows a sustained data transfer rate of 40MB/s. The 30GB is the Deskstar 75GXP (http://www.storage.ibm.com/hdd/desk/ds75gxp.htm) which shows a SDTR of 37MB/s. Seek time for both is 8.5MS, both 7200 RPM, so the benchmark you got might have been for the SDTR. That would account for the difference in speed.

Also, from info I looked up a while back on RAID, in some cases it's really not faster than non-RAID configuration. A search on Google (http://www.google.com) might turn up some useful info. Be prepared though, it took me over an hour to dig through the links and find just what I was after. That's why I didn't do it again...but this is different circumstances, you mgiht find useful info lots quicker...

------------------
Support the right to keep and arm bears.
Note: Please post your questions on the forums, not in my email.

Computer Information Links (http://www.geocities.com/paleopete/)

[RKBA]

Thanks Pete,

Perhaps as you say, the Norton Utilities was ignoring the fact that the RAID array was transferring twice as much data as the single 60GB drive, and just looking at the transfer rate alone. The 37 vs 40 MB/sec looks about right for what NU was reporting (ie; the RAID array was just slightly "slower" than the 60 GB drive). When I get things hooked back up I'll try again and see if I can find a benchmark that shows total throughput and not just transfer rate, because I'd be more comfortable knowing that the RAID0 configuration actually does transfer twice as much data as a single HD.

I can understand how a RAID0 striped array could be less efficient if the two drives were mismatched, but in my case they were identical drives. I spent about a half-hour trying to find something on what you mentioned about it being "not faster than non-RAID configuration" with no luck, but I'll try again when I have more time.

Thanks,

-- Ron

------------------
Centerfire (http://home.pacbell.net/rsdotson/) ‘‘The natural progress of things is for liberty to yield and government to gain ground.’’ -- Thomas Jefferson