i’m glad that worked for you.I experienced the same and went with an aarch64 VM on Oracle Cloud to run my gitlab runners in a docker build environment. A raspberry Pi didn't had enough power to create some of the bigger containers. You can speedup things by using an iSCSI disk instead of MicroSD, yet it's also not that fast, but might work in some environments.
my hope is to find a reasonably performant and local path forward.
I don’t really want to run two different proxmox clusters, if I can avoid it.
I don’t really want to use a bespoke path likely to cause problems for FutureMe…. I have enough of those as it is in other facets of my stacks; and I’d really prefer to keep my hypervisors using official sources…
I don’t want to get stuck in a situation where the fork lags behind were something particularly problematic to arise
I’d prefer to find a way to have aarch64 emulation not suck on x86_64 as that would make things the “cleanest”
barring that, having a way to add alt-arch VM trucks to an existing x64 cluster would be fine.
I can think of a few possible paths for this but I’d really love to hear what the pxm kids are thinking is the way of least pain enduringly, as I’m … moderately confident i’m not the only one desiring something along this vein….
i’d think that a full multi-arch cluster is likely to be painful… as performance characteristics of different arches would likely cause problems with core services trying to migrate from one arch to another feels like a den of papercuts
i’d think that a minimally functional ‘sub-cluster’ or altarch workload nodeset has viability… but i also dunno where dragons are hiding