Decision on 2 SSDs in RAID 1 in Windows 7

[college_student]

Hi all,

I'm hoping to leverage your knowledge to make a decision. For a project at my university, we typically have to load large (5-10 GB) data files (from the SSD) and process them (in visual studio, matlab, etc). I have a latest Xeon workstation with 32GB of RAM. We recently acquired 2 SSDs from a graduating student. Will putting the two SSDs in RAID 1 help speed things up?

Space is NOT an issue at all. So I know by putting them in RAID 1 I'm essentially sacrificing half the total space, but rest assured that space on a single SSD is well above what we need for our application. Will there be a noticeable difference in the RAID 1 setup (when reading large data files in programming applications) over just a single SSD?

Thanks much,
-DV

 

[Essence_of_War]

Raid 1?

I don't think so. Raid 1 is a mirror for redundancy in case of drive failure.

If you put them in a striped raid-0, that should up to double your throughput for large contiguous files.

But that's assuming you're not saturating your SATA throughput...do you know if you have 3 Gb/s or 6?

 

[college_student]

Raid 1?

I don't think so. Raid 1 is a mirror for redundancy in case of drive failure.

If you put them in a striped raid-0, that should up to double your throughput for large contiguous files.

But that's assuming you're not saturating your SATA throughput...do you know if you have 3 Gb/s or 6?

Yes, but with RAID 0 I also increase the chance of failure (if one drive fails, they both fail). I looked online I kept getting conflicting posts of whether RAID 1 actually takes advantage of having identical data on two locations or not.

I have 6Gb/s.

Thanks,
-DV

 

[imagoon]

Yes, but with RAID 0 I also increase the chance of failure (if one drive fails, they both fail). I looked online I kept getting conflicting posts of whether RAID 1 actually takes advantage of having identical data on two locations or not.

I have 6Gb/s.

Thanks,
-DV

It depends on the RAID controller. Some will cycle reads to all drives, others only really use 1 of the disks.

 

[college_student]

It depends on the RAID controller. Some will cycle reads to all drives, others only really use 1 of the disks.

It has the Dell PERC H310. Will this take advantage of the mirroring?

 

[Essence_of_War]

Yes, but with RAID 0 I also increase the chance of failure (if one drive fails, they both fail). I looked online I kept getting conflicting posts of whether RAID 1 actually takes advantage of having identical data on two locations or not.

I have 6Gb/s.

Thanks,
-DV

If you have good, regular, back-ups this isn't as bad as it sounds.

With a raid0, if one drive fails, the raid does fail, but you'll only lose data if you don't have adequate back-ups, and if you have adequate back-ups, whichever drive didn't fail can still be formatted, raid'd w/ a new drive, and then the raid can be restored from your back-ups.

 

[mfenn]

It has the Dell PERC H310. Will this take advantage of the mirroring?

That's an LSI MegaRAID based controller. I couldn't find any evidence that they do load balancing or round robin on reading from a RAID 1.

 

[college_student]

That's an LSI MegaRAID based controller. I couldn't find any evidence that they do load balancing or round robin on reading from a RAID 1.

Oh ok. Would you say that then its a "waste" to RAID these SSDs in RAID 1? If so, then I might just give one of the SSDs to another student in the lab for his workstation.

And I really am not comfortable with RAID 0. Largely b/c I don't backup as often as I should, but that's a whole different issue altogether..

Thanks,
-DV

 

[mfenn]

Oh ok. Would you say that then its a "waste" to RAID these SSDs in RAID 1? If so, then I might just give one of the SSDs to another student in the lab for his workstation.

And I really am not comfortable with RAID 0. Largely b/c I don't backup as often as I should, but that's a whole different issue altogether..

Thanks,
-DV

You'll get protection against a drive failure with RAID1 but that's about it. You should really automate your backup process so that you won't have to remember to do anything.

Another option is to restructure your code so that it can read inputs from multiple directories in parallel.