Datastore Directory Pool Size Allocation issue

[specfiedinfo]

I created a nested 3 HVM nodes to test a local directory pool, created sda (150GB) for boot and os and sdb (10GB) partition for the directory pool feature testing. I have noticed later that the directory pool for data paritition keep changing from the size of the root directory and the mounted directory size. Second parition mounted to /mnt/local_disk, added to the fstab and can be accessed. 

[pkbin]

Hi @specfiedinfo 

This behavior would be seen, if you have not created a filesystem with same size and same mount in all the hypervisor nodes. Since you are creating the datastore ( directory pool) at cluster level, it is expected to have the same mount point in all the partcipated cluster nodes.

Kindly do the below actions to correct it.

1. Remove local_disk datastore  ( infra-> cluster -> storage -> datastore -> remove the local_disk datastore 
2.  Create 10GB partition in all the 3 nodes and mount it as /mnt/local_disk and update the fstab
3. Login back to  VME UI and create a local directory pool datastore by specifying teh mount point "/mnt/local_disk"

Now you must see it correctly 10GB size for the datastore

I hope this helps!

[specfiedinfo]

Hi @pkbin ,

Yes, I have done this steps, the 3 nodes has the same directory name and mount point. at first it show 10GB then suddenly changed to the root size and changed again to 10GB. The issue is it keep changing back and forth showing the correct size of the directoty pool mounted on /mnt/local_disk on each node then changed again to the root size as shown on the image.

[pkbin]

@specfiedinfo  Thanks for this update.

Kindly send me an email  to  hpe-sw-trial-vmessentials@hpe.com ,  We can have a remote session to check on to issue

Alternatively you can provide me the below commands outputs from all 3 nodes
#df -h
#virsh pool-dumpxml  local_disk


Thanks