Windows 10: Intel RST RAID driver in Windows 10 causing disk errors!

So it's a hybrid. The difference is which processor does the work - a PCI board processor or a MB processor. You can't lump it with OS software raid that has no assist from the chip set, which is what software raid implies. It also ports among OSes. I can use my Intel RAID in Linux as well as Windows. My point is you shouldn't generalize like that. It's misleading.

 

I like how no one really responded to my post lol. as far as using a 14.xxx driver as i stated in my post those drivers are incompatible with my chipset.

 

I just wanted to make sure that the very latest one hadn't corrected that.

 

No, the OS driver and the CPU does the work for Intel RST, just like Windows raid does. The only assist provided by the Intel chipset is to make it bootable. Nothing misleading about that.

 

In your own mind you are correct. Have a nice day!

 

You obviously don't understand how it works. Striping and parity calculations are done by the CPU. just like any other software raid. Hardware RAID has dedicated controller hardware that does the striping and parity calculations and can do these much faster than the CPU. Just try to do RAID 5 or 6 on IRST and you'll see how slow writes are because the CPU has to be involved to calculate the parity.

The performance of IRST is the same as other software RAID, such as Windows. That is what matters and that is why it is classified as software raid.

.

 

I understand very well how it works. I also understand that in OS software RAID the driver has no clue about raid - it is only aware of individual drives. All the striping and parity are done by the OS outside of the driver and is not portable across OSes, e.g. Windows and Linux.

In the hybrid such as RST the driver and the chipset set the structure, the driver does the striping and parity not the OS which sees a single disk. A peek in the Windows Storage Manager shows this. That's a significant difference as the OS has no control over the raid system, only the chipset driver, in this case Intel, not Windows. And yes, it's a big difference.

My RST raid 0 is somewhat faster than the Areca 1882 raid 0 for all practical purposes, same drives (2 SSD raid 0), and has proven totally reliable. I used to have it on the Areca controller, but switched to the RST and it works great.

So lets agree to disagree, you feel RST raid sucks and not worth the trouble, and I have had excellent performance and reliability from it. As always, mileage varies. Peace.

 

1. Microsoft Storage Space Controller Driver does the raid functions, not the OS and the OS presents a single disk. Your mixing up Microsoft with the OS.

For RAID 0, the RST and Arecra should be on par - both should scale as the number of drives. It is when you use partity raid that you will start to see the difference,.

Where on earth did you get that I think RST sucks? Software RAID isn't a bad word. I have my two boot SSD and 3 x 1TB spinning disk in RST RAID 0. I wouldn't do that if I thought it sucked.

 

OK, now I hang my head in shame. The remark I thought you made was made by Hydranix. Sigh ...*Doh. My apologees!

But about the driver thing, MS Storage Space Controller is not the driver. It needs to use the MB's SATA controller driver, e.g. RST AHCI or Asmedia SATA or such, to command the individual drives. So there are other software layers whereas RST raid uses a more efficient venue because it is part of the base driver and has intimate knowledge of the underlying architecture and firmware controlling the drives. Which is why you and I are using them.

This type of software raid, if you want to call it that, was slower years ago, but today's sophisticated MB chipset and multi-core processor architecture has closed the gap. For heavy use server use, sure raid controller boards are the way to go, but for personal high-speed computing, not really needed. Which looks like we agree on.

Really sorry about the misunderstanding- I blame old age *Rolleyes!

 

That's OK, I understand that *Biggrin . I am skeptical that it sits on top of SATA drivers because that seems very inefficient and unwieldy. For instance how would spaces handle a drive that goes into a long internal recovery with just the SATA interface. Do you have some reference that Storage spaces uses the SATA driver underneath or other details on their architecture? I would be interested in getting more details on Microsoft's architecture for this,

 

No, I haven't read up on Storage Spaces as I haven't been interested on it other than to understand what it does, but as a long-time software developer it looks only natural to me. Storage Spaces can use any available disk. Those disks are provided by some hardware controller, and the only way Windows knows about them and how to use them is through the drivers provided by the controller manufacturer. It's the nature of the beast. The drivers provide the building blocks Storage Spaces can use.

The fact it can mix and match disparate types of disks make it very flexible and expandable, but also adds to the inherent complexity of the code and thus not as responsive as functionality inherent inside a driver would be. It's the intended trade-off and a very valid one for certain enterprise-level solutions. It's certainly a clever technology.

When you say the drive goes into long internal recovery, what exactly are you referring to? Also remember the disk driver should provide S.M.A.R.T. information which can be used by the higher-level software.

And thanks for being so understanding. I'm still getting over such a dumb slip - they seem to be happening more often, like putting the sugar bowl in the microwave instead of the cup with milk (for a lazy latte) *Rolleyes.

 

Ok i solved my problem with RST software giving me BSOD's. Apparently the Marvel/Jmicron controller created a conflict with the RST software itself not the driver. one of my BSOD's pointed towards Jraid.sys.

To fix this
1. Uninstall RST software in programs and features
2. Go to programs and features remove marvell/jmicron software, restart pc.
3. Download latest JMicron controller for X58 chipset here L: ftp://driver.jmicron.com.tw/SATA_Con...17.65_WHQL.zip
4.Install JMicron controller, Install RST software restart and enjoy full functionality without error

Side note i noticed the RST software changes the way the drives access data and I am not talking about write back cache. After a day or two of the software package running and raid verification the benchmarks improved to a point where it surpassed previous benchmarks without the RST software package installed by a significant amount

Also Windows storage controller is crap. lol at least for the x58.

 

I am a long time software developer as well, including data acquisition and storage, and that don't look natural to me. Now I manage Storage developers and administrators. LOL.

 

A lot of people have said JMicron controllers were crap. I used JMicron for eSATA for a long time and it performed well and worked flawlessly.

RST has much better performance than native windows SATA drivers even for non-raid.

 

Agreed although really i have never tested jmicron drivers not offered by my Manufacturer Asus. From my understanding the JMicron controller on the x58 chipset performs best under AHCI as a single drive and not Raid because it fails to pass trim commands and honestly its older drivers were complete crap, if you look at the changelog the errors they corrected were pretty ridiculous problems that should never have been released in the first place. I just have my DVD player plugged into it.