I'm sure other people have done speed tests and know already that parity drives slow down writes to the array. But I was just experimenting with a speed test with and without a parity drive. I have a gigabit network and transferred over a 4.36GB file.

 

With the parity drive it took 113 seconds. Without the parity drive it took 48 seconds. Pretty big difference, cuts the time down more than half. My peak bandwidth was almost 900 Mbps. In the picture below you can see a visual on the actual test using DUMeter. Just thought it was interesting and figured I'd share it with everyone.

 

Yes, parity slows down the write performance, but not as slow as data loss... That is the point of unRAID.

 

Have you tuned your system?

http://lime-technology.com/forum/index.php?topic=29009.0

 

There's also a feature being considered being re-added back to UNRAID that would allow parity writes to be vastly faster if all drives are spun up. It was taken out in the past as it proved to be a hindrance for larger arrays back in the IDE era.

There's also a feature being considered being re-added back to UNRAID that would allow parity writes to be vastly faster if all drives are spun up. It was taken out in the past as it proved to be a hindrance for larger arrays back in the IDE era.

 

The ability to use reconstruct-write mode was added in 5.0.3.