RAID in discussion
spinner I never said raid-0 was faster then a single disk. I said if you properly setup the same disks as single drives 90 percent of what a desktop user does is no faster including the OS. That means BOTH DISKS. Not one.
But hey man. I mean your the raid guy around here and I bow to your experiance and wisdom as always. You have far more experiance in these matters then us. Looking forward to all your insightful posts as I'm sure we will all learn plenty about disk subsystems and raid systems in general based on reading your previous posts. Talents like yours are unique to say the least. Sorry if I stepped on your toes. Oh Yeah !! Your da man !
Cheers
Tex
But hey man. I mean your the raid guy around here and I bow to your experiance and wisdom as always. You have far more experiance in these matters then us. Looking forward to all your insightful posts as I'm sure we will all learn plenty about disk subsystems and raid systems in general based on reading your previous posts. Talents like yours are unique to say the least. Sorry if I stepped on your toes. Oh Yeah !! Your da man !
Cheers
Tex
0
Comments
I have two RAID 0 systems atm, both Gigabyte/AMD systems, one has 2 x 35Gb Raptors, the other has 2 x WD 120Gb IDE.
The truth is not in benchmarks, but in everyday performance.
Each user 'knows' how good a system feels in everday use. This is not a 'perceived' performance, but in just knowing if a system is working well or not.
I'm no Guru, I'm not a benchmark freak, I've never used SCSI, but I do know what works well and what doesn't.
And in my experience IDE-onboard controller setups work better than a single HD.
I have had two RAID 0 systems go tits-up on me, in both cases a hard disk failed. The first was an IBM 40Gb, the second a WD 120Gb. I was not surprised at the IBM failure, but I was at the WD.
Like I said earlier, I'm no expert, I would not claim to have had vast experience of RAID setups, but I do know what I've used, and I do know you don't have to be a genius to know when a system is performing well.
In my experience, a RAID 0 setup, IDE, controller Card, SCSI, whatever, works better than a non-RAID setup.
But there again, what do I know, eh?
Not a lot
But you said it yourself, (as I did), there is no definitive way of measuring real world performance. So how can you substantiate what you're saying any better than I can substantiate what I'm saying. It's about user preference and opinion, at least when the obvious isn't concerned. We both agree that a single disk is no match for a 2-disk or more RAID 0 setup, because to me, as I'm sure it is to you, that is blatently obvious when IDE is concerned.
I agree that the config you are suggesting, for which I have tried on various occasions, does show benefit, but I still found it was no where near as speedy with regard to everyday desktop activities when compared to a 2 disk RAID 0 setup. That is my opinion, and that is why I responded to the just above quoted comment with "I disagree". I didn't feel what I was saying was offensive or provocative in any shape or form, and I certainly didn't feel it warranted the un-pleasant and combative tone of your following responding post.
With my Raptors? I actually did none of that with those particular drives. I just RAID'ed them straight up. What's your point? Are you insinuating because I haven't tested my current primary rig's drives out in a dual disk Non-Raid setup I'm unqualified to disagree with you? I've tried it your way (at least where the swap file etc is concerned) on previous RAID setups. Like I said above, I felt that the RAID 0 setup was better, by a noticeable amount.
Like I said, I agree there is no definitive benchmark that can reflect priciesly how an OS is going to perform as a result of a specific device group. But I disagree that you can't remotely get an idea for how something is going to perform. I also agree, like I said, there is no one definitive benchmark that can offer anything substantial, but again, like I said, I believe when using a combination of measurements... you can.
Yes, I read what you said, and I thought what you said was sound to an extent. But your resulting conclusion on how that effects desktop performance is something, like I said, I dis-agree with. That is my opinion.
Thanks for that. I appreciate the put down, very big of you.
I know you were of great help to me a few years ago when I was first learning about RAID and all that stuff, I said that. But that doesn't mean anything I say or have learned since then is moot just because I've learnt the rest of it on my own. Or because I don't know as much as yourself.
That's the problem with you Tex, you dis-respect people for what you think they don't know, rather than respecting them for what they do know.
---
Of recent you seem to go out of your way to demean other people and their opinions, I've seen you do it at least a few times with other forum members. On one occasion I recall I even had to step in so as to stop you offending another one of our respected members.
You've even on at least one occasion ignorantly and un-politely seemingly gone out of your way to try and catch me out. Presumably because you thought I didn't know what I was talking about (The recent SP2 news thread). You prowl round the forums with this attitude and it makes a lot of people you respond to feel un-welcome to voice their opinions. That behaviour in my opinion is not in the spirit of these forums.
I don't like it, other folk don't like it, and it turns what could have been a perfectly polite discussion into something like this.
So yeah, I did the un-forgivable thing, I dis-agreed with something you said. Shame on me.
To briefly sumize something with regard to the resulting topic at hand:
I disagree that an IDE dual drive non-RAID setup can be as effective performance wise when compared to a 2 disk or more RAID'd setup using the same disks. I may be wrong, I may be right... but that's my opinion. You obviously can't respect that.
You are right though, I'm no master or guru of drive config's, probably far from it, but I do know a few things. In fact, probably a few more things than you give me credit for. Sufficed to say, the closing comment of your last paragraph pretty much put the final nail in the coffin of our friendship. But I guess you never considered me that did you, not smart enough for that privilege I suppose.
Cheers
From what little I do know about RAID setups is from what TEX taught me years ago...then I promptly forgot most of it. Anyway I beleive that "RAID setups for speed" benefits are most "felt" in applications such as servers where mult...
From what little I do know about RAID setups is from what TEX taught me years ago...then I promptly forgot most of it. Anyway I beleive that "RAID setups for speed" benefits are most "felt" in applications such as servers where multiple, multiple users are accesses the same information/database.
RAID0 setups are also of benefit in audio applications such as studios, where Spinner spends most of his time and video editing suites where I spend most of my time.
Sure the desktop may "feel" smoother but it is the access time reading from the disk that is where the benefit lies. IE: playing back 20-50 tracks or huge chunks of video.
With the game loaded onto my system with a single drive it loads about 40% slower than it does on the same system with a raid-0 array made up of 2 of the same drives.
I run a 32k stripe with one partition on 2-160's and a single 40 gig for my swap file as I stated earlier in the other thread. If I ever reload windows I'll sit down with a stopwatch and try running my apps on one 160 and os on the other with the swap still on the 40, my os and apps on one drive with the swap on the 40 and the array with os and apps on it and the swap on the 40 with all the different setups optimally defragged and see exactly how long it takes for apps like Photoshop, B.F.42, Postal2, GTA Vice City, and all the Office apps to load and document them and try to qualify the real-world benefit to each setup.
Then maybe we can all get along....I have no idea as to when I may do this as my system's tore down ATM in preparation of getting the new water-cooling going and as it's been running super nice as of late so it may well be quite a while.
While I know I'm not one of the more respected members of the forums I do know that when I do throw in my 2 cents worth I'm pretty sure of what I'm saying even if it bucks the trend and while I may not be as learned as some or most of the others in here I'm pretty far from stupid as well.