At least one of the ESX hosts managed by the VC may also show the incorrect storage size value.
3rd party storage device is sending the incorrect storage size values to ESX.
To confirm, follow the below steps to run PCAP from at least one of the ESX hosts included in the datacenter/cluster showing the incorrect value(s).
Generate a packet capture while executing “df -h /vmfs/volumes/<affected_datastore_name>” by following below steps.
Open two terminal (SSH) sessions for the ESX host.
On terminal 1: tcpdump-uw -i <interface> port 2049 -w <PATH>/df.pcap
Note: Change the PATH to some writeable path, and make sure you are not writing the pcap file into the NFS volume.
Example: tcpdump-uw -i vmk3 port 2049 -w /tmp/df.pcap
On terminal 2: df -h /vmfs/volumes/<affected_datastore_name>
Once output is listed from the command on terminal 1, press "CTRL+C", to end the task.
Review the PCAP file and search for "FSSTAT Reply". You should find values for "Total bytes:", "Free bytes:", and "Available free bytes:".
Example:
Total bytes: ########## <<<<< Divided by 1024 "x" number of times will provide the value shown in the vSphere client. If client shows PB value, dividing 5 times provides the value.
Free bytes: ##########
Available free bytes: ##########
If the values listed in the PCAP received from the storage device are not accurate, create a support case with the storage hardware vendor (Pure storage), and work with them for resolution.