I've had the same problem recently. One host in a cluster caused the problem. Although it was showing correct lun size after resizing and rescanning HBAs, something was wrong with it.
Found this out by trying to create new datastore using just added lun space. In the Host selection page of the wisard I selected a host and cliked Next to check if it really could see the lun space. Than clicked Back and select another host. After I found a host that didn't show any devices available, I rebooted it an the issue was solved. Probably, it would be enough to restart some services but i didn't know which exactly.
Personally, I'm afraid that increasing datastore size directly from a host can couse problems in the future.
Make sure that all host accessing that datastore (may be in other clusters) can properly recognize space changes.