2013-08-21 02:57 AM
we've noticed a few times in our Lab, that there seem to be inconsistencies between M&HC and P&C:
Working in a SimLab here, everything Up-to-Date (7-Mode 8.2, Clustered 2-node 8.2, vSphere 5.1U1 (2-node), VSC 4.2).
Sometimes (I've not found a specific rule when) Hosts that appear in M&HC don't appear in P&C.
This is troublesome, since then we can't create datastores with VSC as we have in mind.
I'm trying to push people to use VSC, and this bug doesn't help...
Sometimes deleting the host(s) from M&HC and re-adding them works, sometimes not.
On a similar note, when you create Vservers (VSMs) sometimes it takes a really long time until they appear in the VSC.
Which is funny, because they already are already included in the total numbers in the Discovery Status (of M&HC)...
2013-08-21 07:32 AM
I haven't heard or seen of any discovery issues which would prevent you from provisioning to any existing vServers, nor would prevent you from provisioning to any existing controllers which are visible in VSC.
In the case of the vServers, I would think that using the "update" function would allow them to appear faster for either provisioning or M&HC discovery.
For the hosts/controllers which are not appearing in the M&HC diagram, I would think that we would need to review the VSC logs, with examples of which of the hosts/controllers aren't appearing, including a timestamp at which point they weren't appearing.
Try capturing the following for analysis:
If we aren't able to help quickly over the forums here, we'll go ahead and get you a support case to continue troubleshooting and investigate further.
2013-08-21 09:50 AM
Here's one example from right now:
Note that ISCSIVSM does not appear in the list on the left, it also mentions 1 Vserver...
But in the details of the Cluster LIF, both Vservers appear.
To be fair, I just created ISCSIVMS, but it's been 15 minutes now, and I've hit 'Update' twice already.
Until it officially appears in the list, I can't provision datastores via VSC from it, even though it's mapped and everything.
Could it be, that there's a problem with me trying to implement a Dual-Subnet Design?
It's 30 minutes now, somehow I think it won't appear on it's own... :-(
2013-08-21 11:54 AM
Would you be able to upload the VSC logs? I'm curious to see if there's maybe an inconsistency in the ZAPI calls being sent, or the subnetting is somehow coming into play here (although neither should be true, it will be easier to see in the logs).
In case you haven't generated them before, you can just open your vSphere client, click on "Home" -> "NetApp" -> "Monitoring and Host Configuration" -> "Data Collection", then choose "VSC Logs" and hit "SUBMIT".
You should be able to upload the full bundle to this thread, but if it's too large then ping me offline and we'll get a support case going for you and upload/analyze them from there.
2013-08-21 02:17 PM
Here you go...
On the matter of subnetting there was another quirk:
The 7-Mode System has 4 Interfaces in 4 Subnets configured:
172.16.1.11 -> Mgmt, 172.16.11.11 -> ISCSI1, 172.16.12.11 -> ISCSI2, 172.16.10.11 -> NFS datastores
Only the interfaces on the 11/12 subnet were ISCSI enabled. The System was added to VSC with the Mgmt Interface (e0a).
When trying to deploy an ISCSI datastore via VSC I noticed it was stuck. Checking on the console I noticed that the ESX hosts wanted to log in via the Mgmt interface, which didn't have ISCSI enabled. Only after enabling ISCSI on the Mgmt Interface the provisioning succeeded. Traffic only flows via the ISCSI interfaces as far as I can tell. (That's how I configured it in ESX...)
How does the VSC handle Dual Subnet ISCSI design?
Any known problems?
Greetings from Hamburg/North Germany