Can you enable HA Config on ONTAP Simulator?

[mighty pelican]

Has anyone ever setup two ONTAP Simulators to work like an HA pair? We wanted to test some stuff in a failover situation, but they won't let me set them up in an HA config. I can join them to the same cluster, but can't do takeover/givebacks. Am I doing something wrong, or is this just not allowed?

simulator01::*> cluster show
Node Health Eligibility Epsilon

---

simulator01-01 true true true
simulator01-02 true true false
2 entries were displayed.

simulator01::*> cluster ha show
High-Availability Configured: false
High-Availability Backend Configured (MBX): -

simulator01::*> cluster ha modify -configured true

Warning: High Availability (HA) configuration for cluster services requires that both SFO storage failover and SFO auto-giveback be enabled. These actions will be performed if necessary.
Do you want to continue? {y|n}: y

Error: command failed: Could not retrieve partner SFO information: entry doesn't exist.

[stark plover]

I'm sure it's configurable because CVO and OTS both do it.

Whether it's possible to do so easily is a good question.

[signal holly]

it's not officially supported. There are ways to do it apparently (the NetApp Lab-on-Demand uses simulators in an HA config) but it is neither documented or supported.

But you can deploy an ONTAP Select 2-node cluster with the 90(?) day eval license and use that, that supports HA (and is much faster and can store more data)

[mighty pelican]

Good thought, @signal holly. I will keep that in mind if I ever need to do this again. I was hoping to be able to have our Simulators setup to do this for any testing going forward so that we don't need to setup a whole new environment.

Thanks for the feedback though!

[left sand]

Its quite a bit of work to set up a VSIM HA pair by hand, which is probably the biggest reason its not supported. I've automated it in the past with powershell or ansible. You can see an example in https://github.com/madlabber/labbuilder See the included vsim role: https://github.com/madlabber/LabBuilder/tree/master/roles/deploy_ovf_vsim. OTS HA is much simpler so if that solves it use OTS Deploy.

[signal holly]

that's pretty awesome. You don't happen to have also some details on how to set up a simulator for MetroCluster IP, do you? That would help a lot for our labs (documentation would be fine, doesn't need to be a finished playbook)

[left sand]

I’ve yet to get that to work. The last time I tried it turned out vsim MCC only worked on non promoted builds. But it’s been several years since my last attempt.

[signal holly]

I remember I tried it once by looking at the loader arguments in a lab-on-demand but setting those alone (it was mainly addresses/netmasks for the MC IP interfaces) was not enough I think... but back then I didn't know how to set up an HA pair either which would certainly be a requirement so I didn't try very hard to get it to work

[left sand]

When I last tried it it was a simulated FC style mcc. I got disk shared to all 4 nodes and boot args set but realized there were some probably commands in set test I was going to need. Didn’t get it over the goal line. But IP might be easier.

Let me know how far you get 🙂

[signal holly]

setting these two parameters seems to be enough:

bootarg.mcc.IP_mode="true"
bootarg.ip.if_conf="sw,e0e,e0f"

this hides the two interfaces from ONTAP and puts the MC config in "ready for DR group create". But it still gives some errors during boot that I am not sure about

[left sand]

Sweet. There are more boot args for the 4 way nvram mirroring.

[signal holly]

yeah but those require the cluster UUID which I don't know before creating the cluster ... I'm currently setting that up, we'll see

doesn't look too good at the moment though

it might require a vlan tag (which is tricky)