OS: Antergos (Arch)
MDADM 4
Original RAID 6 array was 8, 4TB drives

I added 4 more 4TB drives (12 total) and the reshape seems a bit slow to me:

active raid6 sdn[11] sdi[10] sdh[9] sdc[8] sdj[1] sdp[4] sdm[7] sdf[6] sde[5] sdd[2] sdo[0] sdk[3]
23441323008 blocks super 1.2 level 6, 512k chunk, algorithm 2 [12/12] [UUUUUUUUUUUU]
[===========>.........] reshape = 56.2% (2199298560/3906887168) finish=1029.3min speed=27648K/sec
bitmap: 0/30 pages [0KB], 65536KB chunk

It's been running for about 18 hours. Reading around on the interwebs, I issued this command: sysctl -w dev.raid.speed_limit_max=500000 (max was set for 200000) and no surprise, this didn't make any difference.

Next up, I was going to try:

echo 4096 > /sys/block/md0/md/stripe_cache_size

(or using the value of 16384 or 32768) but, I'd like to know what the current value is. I thought this was the command:

cat /sys/block/md0/md/stripe_cache_size (yes, I'm a Linux newbie), and I got back: 643 (really odd number....doesn't seem right so I think that wasn't the command I was looking for.)

Lastly, read ahead. blockdev --getra /dev/md0
.....my current value for that is 12288

If it matters, all of these drives are in a SuperMicro 846 with a SAS2 backplane and a single SFF-8087 cable that goes to my LSI 9207-8i HBA. This is a media server in my house.

You have the cat command correct. If you want to dig into the raid modules, install the kernel source and look at the comments in the raid module source files. The speed of reshaping seems about right for the size of the array.

Forgot to write sooner but when it got to about 60-70%, speed jumped up to 80-90MBs. It's all done now. I really didn't do anything though?!

OK, glad it works!