For a machine app that I run, I’ve been trying to clone a vm without success for the past 2 days. I thought it might have been intermittent, since flyctl machine run has always worked for me. The err goes:
➜ fly m clone <vm-id> --config <toml> --name <machine-name> --region <r>
Oops, something went wrong! Could you try that again?
The problem with run is, it doesn’t replicate env vars. I am not sure if clone does, but I’d prefer clone if it does (but it seems like clone and run are one and the same thing and that clone isn’t special in any way, shape, or form).
Is fly m clone going away? And never supposed to work?
cc: @eli clone still wouldn’t work… is this expected because the preview announcement mentioned it should?
May be it is reserved for the upcoming migrate-vm capability?
Hi @ignoramous. Why is this thread marked as solved? Is flyctl m clone supposed to work now? I still get the same error when trying to clone:
Oops, something went wrong! Could you try that again?
I am using clone to scale a postgres app. If clone does not work, what is the recommended way to scale it?
Hi there,
Check if you are running the latest flyctl version (docs)?
See what LOG_LEVEL=debug flyctl m clone ... tells you?
Alternatively, if you are up for it, you can checkout, compile flyctl, and run the resulting binary, you should see failing commands echo out a full stack trace (prev). You could then report the issue back upstream.
The issue I faced was with regular machines which was subsequently fixed, and not with Fly Postgres v2.
I updated flyctl to v0.0.445 yesterday and was seeing the issue with this version. This morning, new update was available again on MacOS, so I updated to v0.0.447 and this time the cloning works. 
Thank you @ignoramous.
1 Like