I have a node that was deployed from a MAAS server. Sadly, the node survived the MAAS server and now resides all alone.
It works, but it takes a lot of time to do a reboot.
I’ve been looking, but I can’t find any documentation on HOW to remove the MAAS components that are still on the node. Is there a way to do this?
BR/Patrik
@pal-arlos, i’m not sure we have doc on how to remove MAAS components from a node; the expectation is that they will be burned down and reloaded, either by MAAS (if they’re still managed there) or by hand, (if they aren’t). so i think the answer is kinda, “no”, there’s no (easy?) way to do that.
but i have to ask: you said, “Sadly, the node survived the MAAS server and now resides all alone.” if that’s the case – if you didn’t expect it to survive MAAS, why not just reload it? have you dropped MAAS completely?
@billwear It’s been a while now, so I try to recall as much as possible.
We started out with one MAAS (M1), from it we deployed a ‘clean’ host H1, and all was well. The host was working fine, and doing its job.
I can’t remember if it was a hardware failure on the M1 node.
Anyway, the result was that we decommissioned (fresh-install) the M1 node, and deployed a new MAAS (different device, different MAAS version) M2.
However, as H1 was still operating and we had no redundancy on the service it provided, we could not shut it down, and move it to the control of the M2. Hence, H1 was left without a MAAS.
We’re keeping MAAS and expanding the machine base it is managing. But we will probably also have a parallel PXE system, that can deploy stand-alone hosts that will not depend on cf. MAAS or other boot-strapping systems. I…e a system that will deploy MAAS rackd and regiond.
BR/Patrik
yeah, that failure baffles me, @pal-arlos, but it sounds like you’re doing responsible design. distrust all claims for the ‘one true way’ 
@pal-arlos,
From what I’ve seen, MAAS sets the flags in the bios to PXE boot on it’s network card, then adds a few configuration settings for things like Name Resolution and Proxy services.
Undoing them isn’t impoosible, but would be a learning experience. I’ve done part of that on one host my MAAS started, and it’s booting much faster after tweaking the BIOS to boot from the hard drive on the host.
That may at least resolve your slow booting problem.
Good luck - and perhaps post back here once you’ve gone through the host with a fine-tooth comb looking for issues to fix your MAAS’lesness
Hope that helps
Thanks
~~ Charles
@pal-arlos, yeah, i think you’re probably going to have to create a redundant H1 node (H1a?) and turn off H1 to recover it. after playing around with things, it feels like it can’t know that M1 has gone completely away until you power it down and recommissioning it with M2.
@billwear, would have been nice with a ‘non-destructive’ commission, i.e., bring in an existing host into the MAAS herd.
But, until that exists, I’ll just hope for the best, and that nothing bad happens with H1. In my view, this is the drawback of MAAS. Hopefully, it’s a non-issue, if running HA.
@pal-arlos, thanks for this feedback. you should describe what you want as a feature, and add it to the features category. carries more weight when it comes from an actual developer!
