intel Matrix Raid/Rapid Storage Technology

[glennd]

is intel Matrix Raid/Rapid Storage Technology ok? I realise it's software raid but does it work ok? Any known issues? There is a customer machine about to come under my jurisdiction with 2 x 240GB SSDs hooked up to this. What should I be looking for, watching etc?

 

[phaZed]

It seems to work fine most of the time. There have been countless times, however, that new RST drivers or seemingly unrelated system drivers were updated that broke RST and causes a bluescreen on boot. Getting the system back after this can be hit or miss. Not exactly what you want to look forward to while using a RAID array.

Is this system going to be RAID 0 or 1? I would recommend against RAID0'ing any SSD's - it's generally not an improvement or only slightly an improvement in speed for SSD's. Hard drives, sure.. SSD's, not so much. If your RAID1'ing the drives then that's fine.. if RST screws up, at least you can work on simply one drive and it's not striped or reliant on more than itself. Turn off raid, boot it like a normal drive and your back.

 

[glennd]

It seems to work fine most of the time. There have been countless times, however, that new RST drivers or seemingly unrelated system drivers were updated that broke RST and causes a bluescreen on boot. Getting the system back after this can be hit or miss. Not exactly what you want to look forward to while using a RAID array.

Is this system going to be RAID 0 or 1? I would recommend against RAID0'ing any SSD's - it's generally not an improvement or only slightly an improvement in speed for SSD's. Hard drives, sure.. SSD's, not so much. If your RAID1'ing the drives then that's fine.. if RST screws up, at least you can work on simply one drive and it's not striped or reliant on more than itself. Turn off raid, boot it like a normal drive and your back.

Raid 1. To add some redundancy to the system along with the backup regime. It sounds like once the system is running and stable, we want to discourage windows from updating any drivers. I would expect if the system updates the drivers, that will get mirrored to the second drive so if it breaks I'm hosed anyway.

Now for my education: When the main drive fails, at what point and how exactly does that drive get demoted out of the system and the second drive kicks in. Is it a case of I have to reboot and let the system detect the failed drive and boot into the second drive?

 

[phaZed]

Now for my education: When the main drive fails, at what point and how exactly does that drive get demoted out of the system and the second drive kicks in. Is it a case of I have to reboot and let the system detect the failed drive and boot into the second drive?

If the drive fails, RST will actively demote it and detect issues in real time, indicated by a Yellow Exclamation mark in the software and system-time area popups. The system will survive a drive failure or having a single drive unplugged while running... but your board must support HOT PLUGGING, in the case of physically unplugging. If a bad drive is found, shut the system down, replace the failed drive with a new, blank drive and start the CPU. The array rebuild process is done, again, from the RST software in Windows.

I have never had success of hot plugging a system of it's bad drive, swapping in the new, and then performing a build in the same OS session... a reboot seems to be required on the systems I worked on. So, if it were like a mission-critical server and you wanted Zero-downtime for drive swaps... RST isn't the right choice.


I may sound like I'm crapping all over RST, but it's really not bad and has progressed a lot in recent years. In your RAID 1 case I would be personally comfortable with using it.

If a drive fails, the computer will still work normally and boot after said failure using a single drive, automatically.

IMO, RAID1 is the only raid to run on RST - hardware for everything else.

 

[glennd]

If the drive fails, RST will actively demote it and detect issues in real time, indicated by a Yellow Exclamation mark in the software and system-time area popups. The system will survive a drive failure or having a single drive unplugged while running... but your board must support HOT PLUGGING, in the case of physically unplugging. If a bad drive is found, shut the system down, replace the failed drive with a new, blank drive and start the CPU. The array rebuild process is done, again, from the RST software in Windows.

I have never had success of hot plugging a system of it's bad drive, swapping in the new, and then performing a build in the same OS session... a reboot seems to be required on the systems I worked on. So, if it were like a mission-critical server and you wanted Zero-downtime for drive swaps... RST isn't the right choice.

Sounds like it will work for this case. In the event of failure, running on the second drive gives me the breathing space to organise the replacement drive. zero downtime is not required, i can go onsite after hours to bring down the server and replace the drive.

I may sound like I'm crapping all over RST, but it's really not bad and has progressed a lot in recent years. In your RAID 1 case I would be personally comfortable with using it.

If a drive fails, the computer will still work normally and boot after said failure using a single drive, automatically.

IMO, RAID1 is the only raid to run on RST - hardware for everything else.

It's good to know what it's limitations are so I don't get caught out.

Thanks for your input.