I have 2 VMware ESXi hosts. I have successfully created and connected an iSCSI (IBM DS3300) storage array to the first host. I have created a VM on that datastore and it is running just fine. When I go to add that datastore to another host I go to Configuration\Storage Adapters and click rescan. It detects the iSCSI array. I then go to Configuration\Storage and do not see the datastore, so I go to add it, it then sees the iSCSI array, but tells me "Warning: The current disk layout will be destroyed. All file systems and data will be lost permanently."
Don't use 'add storage', this will re-format the volume! It's probably not mounting the VMFS on rescan because the LUN ID of the volume is not the same on both ESXi hosts. This will cause the second to see it as a snapshot LUN and refuse to mount it. Check your DS3300 config and make sure the volume is presented with the same LUN ID to both hosts, and then try another rescan.
Alex
It is presented as the same LUN. My DS3300 is configured with a host group called VMFS. Both hosts are members of the group. The LUN is pointed at the host group. Both hosts see this as LUN 0.
What else can I check?
So have you done a rescan in Configuration \ Storage Adapters on the 2nd host?
Are your ESXi's in the cluster or standalone hosts?
---
VMware vExpert '2009
@Dave - Yes, have rescanned numerous times. I also built a new (stand alone?) host, added it to the host group and it has the same issue, it want's to format the datastore.
@Anton - Not sure what you mean by cluster or stand alone. My hunch is stand alone, but my hardware is an IBM BladeCenter S with two blades. We are not running anything other than ESXi, no VMotion, no VServer, etc., just the plain jane entry level hypervisor. I believe them to be considered stand alone even though they are in the same blade chassis.
Yeah, they're standalone and will stay standalone unless you set up vCenter and combine them to cluster.
---
VMware vExpert '2009
They can still share an iSCSI volume even if they're stand-alone hosts though, no?
Do you see any errors in the logs for ESXi after you do the rescan?
You can still share the LUNs with stand alone hosts.
SCSI: 4506 Cannot find a path to device vmhba33:0:3 in a good state. Trying path vmhba33:0:3. (2:01:12:45:593 cup3:590665)
That is followed by a number of:
LVM: 4476: vml.(long hex string?):1 may be snapshot: disabling access. See resignaturing section (2:05:40:21.831 cpu0:1836)
That error is consistent with the LUN ID being presented to the host is not the same as the LUN ID of the volume on the host it was formatted on. Check again that both hosts are presented the volume with the same LUN ID.
Also looks like you've got a path to the volume down so check on that too....maybe you're presenting it with different LUN IDs to different paths.
Alex
Thought I'd post back the solution to this.
Open the VI Client
Go to Configuration
Go to Advanced
Go to LVM
Set LVM.DisallowSnapshotLun to 0