Richard Siemers wrote:
When you "export" a VV(set) to a host(set), what is really happening is a VLUN template is created. So the concept of "unexport" is dangerous because its misleading as it is really deleting a VLUN template.
If your VVSET was exported to each cluster node, 1 at a time, then you should be able to see several VLUN templates, find the one for the host you want to nuke, and delete that 1 VLUN template in the IMC.
I think it is possible to export a VVSET to several hosts at once, under one VLUN template, which would paint you into the corner I think you are describing. Double check your VLUN templates to see if its 1 per host, or one big VLUN.
Looking at the VVs in our VVsets, I see that a VLUN template has been created for each host that it is exported to. So, manipulating a VLUN template and the associated active VLUNs on a host should not impact the underlying VV for any other hosts, right?
I have noticed that VVs and VVsets exported to a Host Set don't appear to have individual VLUN templates per host. But they do have Active VLUNs on that host, which if I purged those for the host in question I should be okay. Of course in the VVset situation, ideally I'd just edit the VVset and remove the VV.
In the end, though, if export is to create a VLUN template and active VLUNs, then an unexport is effectively removing of the active VLUNs and, but not always, the VLUN template, depending on how the VV or VVset was exported to the host.
How common is it for people to manipulate VLUN templates and Active VLUNs directly? It doesn't seem like that big of a deal, as long as you connect the dots of what you are doing, which you'd do anyway before messing with exports.