Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am doing a resync on a RAID6 array and I am looking at some 2 days to complete. I looked here for hints on how to speed it up, but I have, instead, managed to slow it down!

HW details are: 10x 4TB WD Black HDDs (one died so resync is due to this) Core2Quad, ASUS P5K Deluxe with 8GB RAM

What are the best params (VALUE) to use for this config and/or how do I calculate these?


echo VALUE > /proc/sys/dev/raid/speed_limit_min (I set to 90000)
echo VALUE > /proc/sys/dev/raid/speed_limit_max (I set to 900000)

Hint 2 - READ AHEAD:

blockdev --setra VALUE /dev/md0 (set to 32MB / 65536) 

Hint 3: STRIPE-CACHE SIZE (how do I find system_page_size? I set to 32MB / 32768 and have not seen OUT OF MEMORY error):

memory_consumed = system_page_size * nr_disks * stripe_cache_size 
echo VALUE > /sys/block/md0/md/stripe_cache_size 

I have disabled NCQ and enabled internal bitmap before starting resync so these steps are working.


share|improve this question
On high spec drives there is no benefit to disabling NCQ – JamesRyan Mar 17 '14 at 8:25
Hi JamesRyan, hmmm the resync increased from 35K to 40K by doing this. It is the rest that has had little effect. Thanks for the reply. – MarcusL Mar 17 '14 at 9:28
That implies that the drive isn't your bottleneck then. – JamesRyan Mar 17 '14 at 9:52

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.