Bug 948745

+++ This bug was initially created as a clone of Bug #903212 +++

Description of problem:
During expanding the size of RAID1 or RAID5 volume the resync process is reported in /proc/mdstat but there is no information about it in the volume’s metadata. Therefore if the RAID volume was stopped during the process of size expansion all information about this progress would be lost, because no checkpoint is saved in the metadata.

Version-Release number of selected component (if applicable):
mdadm-3.2.6-14.fc18

How reproducible:
Always

Steps to Reproduce:
1. Create RAID1 or RAID5 volume of size less than maximum and wait for its normal state, e.g.:
# mdadm -C /dev/md/imsm -amd -e imsm -n 3 /dev/sda /dev/sdb /dev/sdc
# mdadm -C /dev/md/volume -l5 --size 10G -n 3 /dev/sda /dev/sdb /dev/sdc

2. Expand the size of a RAID volume providing new size, e.g.:
# mdadm -G /dev/md/volume --size max

3. Check metadata during rebuild process:
# mdadm -E /dev/md/imsm
  
Actual results:
Metadata contains no information about the resync process ongoing during expanding the size of RAID1 or RAID5 volume. No checkpoint is saved. Metadata shows normal state of the volume.

Expected results:
Metadata contains information about the resync process ongoing during expanding the size of RAID1 or RAID5 volume. Checkpoint is saved.