VMware Cloud Community
esx-latrobe
Contributor
Contributor

VC 2.0.1 shows incorrect datastore free space

We have a 2TB LUN presented to 13 ESX hosts (various versions of 3.0.0, 3.0.1 and 3.0.2). When I do a vdf -h on all the hosts I get the same (presumably) correct result:

Server: esx---

Filesystem Size Used Avail Use% Mounted on

/vmfs/volumes/472fbcaf-7eaff269-7b30-0017a4770000

2.0T 1.4T 580G 71% /vmfs/volumes/472fbcaf-7eaff269-7b30-0017a4770000

The problem is, when I look at the same datastore information in the VC client via VC server it is reported as empty (2TB free) on all hosts. The really strange thing is, when I go to Configuration -> Storage on all the hosts and select the datastore, in the Details section on all it says 2TB capacity, 710MB used, 2TB free, which just doesn't add up.

If I connect my VC client directly to the hosts I get different results again, two report accurate info (matches what vdf tells me) but the others do not. Four report the datastore empty (well, 710MB used, 2TB free - which matches what VC server sees) the rest report various states of free, none of which match anything else.

Any clues as to what's happening here? I've tried Refreshing the storage view in VC server and also with VC directly on the hosts - makes no difference. I'm pretty sure the datastore was formatted on a freshly installed 3.0.2 (update1) host - would that make a difference? The other 3.0.2 hosts were updated to Update 1 via esxupdate.

Tags (1)
0 Kudos
1 Reply
MrBunjit
Contributor
Contributor

I've just discovered the same problem today, luckily on a smaller setup - here's how I fixed it.

Set up is two ESX 3.0.2 servers, VC 2.0.2, 250GB iSCSI SAN volume on Equallogic.

Procedure:

Created volume on SAN

server 1 - iSCSI rescan

server 2 - iSCSI rescan

server 1 - format volume

moved 4 vPCs onto volume 8GB drives each

Result: VC sees 249GB free, connecting directly to server 1 sees 215GB free, server 2 sees 249GB free.

Fix: did a iSCSI rescan on server 2 and now all servers and VC show 215GB free. Seems it was caused by rescanning for the volume on server 2 BEFORE it was formatted.

Hope this helps...

0 Kudos