Created attachment 160891 [details] dmesg, sysrq+t mdadm-3.3.2-1.fc21.i686 REPRODUCE STEPS Create four devices to use: 1. for i in {1..4}; do fallocate -l 1073741824 "file$i"; done 2. for i in {1..4}; do losetup -f "file$i"; done Create array: 3. mdadm -C md0 -n 3 -l raid5 /dev/loop[012] 4. cat /proc/mdstat Personalities : [raid6] [raid5] [raid4] md127 : active raid5 loop2[3] loop1[1] loop0[0] 2095104 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU] Add+Grow array: 5. mdadm --add /dev/md/md0 /dev/loop3 mdadm: added /dev/loop3 6. mdadm --grow --raid-devices=4 /dev/md/md0 mdadm: Need to backup 3072K of critical section.. Hang. dmesg shows: [ 1043.523780] RAID conf printout: [ 1043.523857] --- level:5 rd:4 wd:4 [ 1043.523918] disk 0, o:1, dev:loop0 [ 1043.523978] disk 1, o:1, dev:loop1 [ 1043.524087] disk 2, o:1, dev:loop2 [ 1043.524148] disk 3, o:1, dev:loop3 [ 1043.534806] md: reshape of RAID array md127 [ 1043.534871] md: minimum _guaranteed_ speed: 1000 KB/sec/disk. [ 1043.534928] md: using maximum available idle IO bandwidth (but not more than 200000 KB/sec) for reshape. [ 1043.535063] md: using 128k window, over a total of 1047552k. [ 1043.739359] md127: unknown partition table And after ~10 minutes, no progress on reshape: # cat /proc/mdstat Personalities : [raid6] [raid5] [raid4] md127 : active raid5 loop3[4] loop2[3] loop1[1] loop0[0] 2095104 blocks super 1.2 level 5, 512k chunk, algorithm 2 [4/4] [UUUU] [>....................] reshape = 0.0% (0/1047552) finish=19.5min speed=864K/sec dmesg still doesn't show any issue. So I've issued a sysrq+t and will attach dmesg.
Created attachment 160901 [details] mdadm -D
Created attachment 160911 [details] mdadm -E
Changed to 64-bit hardware, same results, reshape hangs. kernel-3.17.6-300.fc21.x86_64 mdadm-3.3.2-1.fc21.x86_64