Disk resize not applied at instance level

Issue

• 2 weeks back some multipath disks got resized successfully according to Openstack and multipath. When checking the instances, we noticed that the disks did not got resized for 1 out of the 4 resized disks (1.5TB to 2TB) on 1 VM. And 0 out of 4 on the other VM.

• Openstack output:

| f1cbd067-0f7b-4d3b-880d-cbdf8948bf6d | oraclecl2db-datafiles2_ldg             | in-use    | 2048 | cinder_rhos_prod_ldg | false    | Attached to guestVM2.localdomain on /dev/vdp Attached to guestVM.localdomain on /dev/vdp  |            |
| 2c9cc7bc-0258-408a-a33a-44697a2c3666 | oraclecl2db-datafiles2_thr             | in-use    | 2048 | cinder_rhos_prod_thr | false    | Attached to guestVM.localdomain on /dev/vdo Attached to guestVM2.localdomain on /dev/vdo  |            |
| b5ecadbe-5ac1-4dc4-90fb-391f6a7ff642 | oraclecl2db-datafiles1_ldg             | in-use    | 2048 | cinder_rhos_prod_ldg | false    | Attached to guestVM2.localdomain on /dev/vdn Attached to guestVM.localdomain on /dev/vdn  |            |
| 6535ec16-be89-4faa-8253-fe69bf07e63a | oraclecl2db-datafiles1_thr             | in-use    | 2048 | cinder_rhos_prod_thr | false    | Attached to guestVM2.localdomain on /dev/vdm Attached to guestVM.localdomain on /dev/vdm  |            |

• The disks got resized according to the Openstack output.

-Multipath output:

000000000000000000000000000176 dm-83 NETAPP,LUN C-Mode
size=2.0T features='2 pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| |- 1:0:2:42  sdka 65:480  active ready running
| `- 2:0:3:42  sdkb 65:496  active ready running
`-+- policy='service-time 0' prio=10 status=enabled
  |- 1:0:1:42  sdjz 65:464  active ready running
  `- 2:0:1:42  sdkc 66:256  active ready running


000000000000000000000000000175 dm-81 NETAPP,LUN C-Mode
size=2.0T features='2 pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| |- 1:0:2:41  sdie 134:224 active ready running
| `- 2:0:3:41  sdif 134:240 active ready running
`-+- policy='service-time 0' prio=10 status=enabled
  |- 1:0:1:41  sdid 134:208 active ready running
  `- 2:0:1:41  sdig 135:0   active ready running


000000000000000000000000001443 dm-85 NETAPP,LUN C-Mode
size=2.0T features='2 pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| |- 1:0:3:42  sdke 66:288  active ready running
| `- 2:0:5:42  sdlq 68:384  active ready running
`-+- policy='service-time 0' prio=10 status=enabled
  |- 1:0:4:42  sdkd 66:272  active ready running
  `- 2:0:4:42  sdlp 68:368  active ready running


000000000000000000000000001442 dm-82 NETAPP,LUN C-Mode
size=2.0T features='2 pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| |- 1:0:3:41  sdjw 65:416  active ready running
| `- 2:0:5:41  sdjy 65:448  active ready running
`-+- policy='service-time 0' prio=10 status=enabled
  |- 1:0:4:41  sdil 135:80  active ready running
  `- 2:0:4:41  sdjx 65:432  active ready running

• So, according to multipath, all 4 disks got a resize. Same on the other physical host.

• The instances output of the first VM:

[root@guestVM (prod) ~]# lsblk
NAME               MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
...
vdm                251:192  0  1,5T  0 disk 
vdn                251:208  0  1,5T  0 disk 
vdo                251:224  0  1,5T  0 disk 
vdp                251:240  0    2T  0 disk 
...

• The other VM in front of the multipath LUNs has all 4 disks showing 1.5T. So no resize captured here.

• What I did notice is the following:

()[root@overcloud-compute-0 /]# virsh domblkinfo instance-0000ad10 vdp --human
Capacity:       2.000 TiB
Allocation:     1.153 TiB
Physical:       2.000 TiB

()[root@overcloud-compute-0 /]# virsh domblkinfo instance-0000ad10 vdo --human
Capacity:       1.500 TiB
Allocation:     1.153 TiB
Physical:       2.000 TiB

()[root@overcloud-compute-0 /]# virsh domblkinfo instance-0000ad10 vdn --human
Capacity:       1.500 TiB
Allocation:     1.153 TiB
Physical:       2.000 TiB

()[root@overcloud-compute-0 /]# virsh domblkinfo instance-0000ad10 vdm --human
Capacity:       1.500 TiB
Allocation:     1.153 TiB
Physical:       2.000 TiB

• So, according to virsh, the physical size is 2TB but for some reason the logical / vm disk size is still 1.5TB for 3 out of the 4 resized disks. On the other VM, all 4 have the 2TB physical size as well but all 4 also show the logical / vm disk size of 1.5TB.

• Is there an online possibility to fix the Capacity?