• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

What Raid are good?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Dude - sorry but you are simply way off base. Show me some HDTach graphs of a 4 disk RAID0 from onboard controller that even remotely supports what you are trying to say. 15% or more is not uncommon for this test! 1% is very UNCOMMON unless you have onboard processing on your RAID card- and even then, that is very low.

The problem is that you're limiting yourself to Windows and/or onboard softRAID controllers. I'm not going to repartition my home system just to prove you wrong.
 
Originally posted by: Nothinman
Dude - sorry but you are simply way off base. Show me some HDTach graphs of a 4 disk RAID0 from onboard controller that even remotely supports what you are trying to say. 15% or more is not uncommon for this test! 1% is very UNCOMMON unless you have onboard processing on your RAID card- and even then, that is very low.

The problem is that you're limiting yourself to Windows and/or onboard softRAID controllers. I'm not going to repartition my home system just to prove you wrong.

What are you talking about? 😕

 

A typical 4-disk RAID0 in HDTach can use up to 15%, or even
more sometimes, of the CPU for a typical onboard, or other software RAID solution. In a RAID5, the CPU utilization would be SLIGHTLY higher, but still VERY low.

EDIT: Misread, and someone totally missed that he was speaking explicity about software RAID, of which I have no experience with... Here are results from a dedicated RAID card though....
Tas.
 
Originally posted by: tasburrfoot78362

No, you are incorrect. A typical 4-disk RAID0 in HDTach can use up to 15%, or even
more sometimes, of the CPU for a typical onboard, or other software RAID solution. In a RAID5, the CPU utilization would be SLIGHTLY higher, but still VERY low.

Um... Sorry, no. And this benchmark proves it.... And RAID 5 is considered to be require more intensive calculations.
Tas.

Did you actually READ what I wrote?!?!?!

I'm talking about a SOFTWARE raid solution - you know, a crappy card or onboard solution.

You showed me a good quality card with dual Xeon setup.. how does that show I'm wrong.

And btw, in mentioning RAID5- I'm talking about slightly higher utilization in a GOOD QUALITY card- and your bench proved it. You still have less than 5% utilization.

Thanks for proving what I'm saying about good quality stuff.

EDIT: and here's my HDtach score: RAID 0: 656.5mb burst / 192mb average @ 1% CPU use (+/- 2%)
 
Originally posted by: Cr0nJ0b
remember that even with mirroring raid 0 you can still lose ALL of you data if you get a filesystem error or a virus. backups are the only way to ensure data protection.

Having used NTFS since '93, I've only experienced filesystem corruption once (faulty motherboard/memory corrupted the IO), but hard disk crashes have been numerous. (curiously enough, at home I use SCSI and never had to replace any drives, but at work we buy replacement IDE&SATA drives all the time)

Viruses are equally rare. (I receive them all the time of course, but never activate them)

But yeah, backups are obviously important! The thing is that restoring the backup after a drive crash takes time. Where I work we have several servers that run our own software and do not really store anything, but if the drive crashes the thing will crash too. RAID-1 allows us to replace one of the drives while the server remains up&running. (in short: Lots of time saved)
 
What are you talking about?

That software RAID0 should use virtually no additional CPU time. I mean hell, the I/O request handler for the md RAID0 driver in Linux is less than a hundred lines of C. If the Windows software RAID0 driver causes 14% overhead something is seriously wrong with it.
 
Originally posted by: Rilex
RAID IS NOT BACKUP. RAID is for Redundancy.

Period. End of story.

If your "backup" is on a platter, it is not a backup.

For a typical home user, maybe.

For the enterprise, you could not be more wrong...

Edit for symplicity in a home setup. Any IO intensive app should have it's own spindle. Otherwise you will put in unnecessary amount of RAID overhead. Running multiple IO intensive apps on spindle is what can easily turn a computer into a sluggish beast. You could branch out and running a multi-spindle system, have apps that take advantage of better read/write performance. For the home setup, like here, RAID is adding in a layer of complication that you could avoid by keeping each disk for a discrete purpose.

Linky for general RAID info here
 
Back
Top