Hi, On Sat, Jun 01, 2019 at 10:50:24AM +0200, keld@xxxxxxxxxx wrote: > Still, Andy, you need to cover all layouts of md raid10. > > L know that for the far layout we actually had something that meant choosing the faster drives > an thus it violated the striping on HDs, degrading read performance severely. A patch fixed that. > > this patch did not apply to the offset layout, so maybe that layout could satisfy your needs. Okay. Which layout combinations are you interested in seeing results for? Obviously I've done 'n2' already as it's the default, so is it just 'f2' and 'o2' that you would be interested in? I don't think there is any point in changing the number of copies from 2, do you? Is it enough to stick with a random read test (best at exposing the differences we are talking about in this thread) or would you like to see the other tests like sequential read/write and random write too in case there are other differences? Finally, do you consider 4KiB IO sufficient? RAID-10 should allow striping and there is not much opportunity for that with 4KiB IO and a default 512KiB chunk size. What I can't do is test more than 2 devices nor equal performance devices (unless you are interested in what happens with two 5.4kRPM HDDs, that is), because all I have in a test host right now is the SSD and the NVMe. Cheers, Andy
