Whats the correct way to resize iscsi gfs2 datastores?

[BMuellerDSR]

Hi to all,

whats the correct way to resize an shared iscsi gfs2 datastore?

What I did:

Resized the iscsi volume at the storage system

Rescan iscsi targets on one host (after Resize)

iscsiadm -m node -R

sudo systemctl restart multipathd

gfs2_grow /mnt/7e2728f1-47de-4832-8893-2ac90e7c9451

On all other Hosts

iscsiadm -m node -R

sudo systemctl restart multipathd

After that all hosts now show the volume with the right new size via LSBLK

(I grew the gfs2 volume from 60GB to 90GB)

But on the webif of vmemgr it shows complete nonsense, Two hours it showed a datastore size of 2.1MB!

After Restart of the vmemgr and all three hosts it is showing a new size of 67GiB (more than 60GiB but much less than 90GiB) how to get the system to show the right size? And whats the right way to grow gfs2 datastores?

Thank You.

[PeterTzvetanov]

1. Created a LUN and add this one to the Cluster as second GFS2 datastore.
2. After that I extended it on the storage level with 100GB.
3. I run a rescan across all ubuntu hosts, but still couldn't see the new size reflected.
4. I set the one host in maintenance mode via the VME and via ssh set the node in maintenance via "pcs node standby" and I rebooted it.
5. Once host came back I run "pcs node unstandby" and lsblk give me the new size, but the GFS2 was still on the old one.
6. I followed up command multipathd -k'resize map 360002ac000000000000000890002d515'
7. Then the gfs2_grow -v /mnt/1fa3bf99-4efb-4423-9fca-c31339ee7c00
8. Now the first ubuntu host saw the new size, but I had to reboot one by one the others and always first set particular host in mainenance mode as in step 4 and 5. After that each rebooted host over the df -h I could se the new size.
9. Once I finished with all ubuntu hosts in the cluster, I did a cluster refresh/sync manually and once that was finished I could see that the size that is expected reflected also in the VME GUI.