<<
I would say that ALL IDE drives are programmed to optimize thoughput and acess time, so that wouldn't make any difference... Spikes in throughput will be made when necessary and is not limited to Maxtor drives... >>
Yes they are optimized in their own way. However, some manufacturers will tune these optimizations for a particular benchmark, for example. This timing causes the transfer aberrations as seen in WB99 disk inspection test; transfer.
<<
I have heard all of this before, i have set up systems with mixed drives in RAID 0 and RAID 5 arrays just to see if anything anyone said about it would happen, because it did not make any sense to me... >>
What if anything did you use to test the results? Read/Write speed amongst various block sizes and que depths?
<<
Well guess what, you still cannot explain it properly, and i still have no problems with my mix of Maxtor, quantum and fujitsu drives (on a Adaptec IDE RAID card), and i am pretty sure they use different firmware... ;-) >>
Just because it works doesn?t mean it?s working they way a user would want it to in order to see the benefits of doing so in the first place!
<<
I also use a combo of two Atlas 10K coupled with two Seagate 15X drives in a RAID 5 array without problems, have been for the last year... >>
Apples and Oranges! We are talking about ATA devices. SCSI devices must adhere to strict ANSI guidelines. The chances of things going extremely awry are far less here.
<<
Different speeds, firmwares, interfaces (as long as they are backwards compatiable), brands do not affect the stability of a RAID 0 array, it is the actual RAID 0 array that doubles the risks... >>
I?m not talking about the overall stability of an array. I?m talking about it working like it should; a user that wishes to stripe two or more drives is looking for increased read and write rates. They are not concerned with fault tolerance. If they are as concerned about fault tolerance as they are with STR, they would use RAID 50 and ECC!
<<
The rumours about different drives being better than others when used in RAID 0 is turning into an urban legend... >>
These are NOT rumors but fact! There are hundreds of users having the same problem that goes back to mid 1999. The problem first showed up with Maxtor disks. If you go over to Storage Review?s messageboard, you will see what I?m talking about!
The problem (outside of any other problems like ?it doesn?t work!?) is across the entire volume, the STR will indicate strange patterns in the graph. These are unlike the downward spikes seen in HD Tach where the OS is accessing the tested volume. WB99 inspection runs in realtime to reduce the chance of interference from other processes. Even so, the ?problem? drives would show a very repeatable ?zig-zag? pattern which resembles a stitch produced by a sewing machine.
If you assemble systems for people your main concern is to get it working stably. Of course, most users that are interested enough to stripe their disks for better R/W performance are going to have the know how to test it with familiar and downloadable software. This is where these difficulties emerge.
Cheers!