[-] maarvin@lemmy.world 3 points 1 month ago

Proxmox with all of its warts appears to be the better hypervisor. If your host has the headroom like others suggested: proxmox with a VM hosting docker/podman.

I have to say running proxmox saved me from running to the basement or opening up IPMI a few times already when making suspect changes to a VM that otherwise would have taken down the network of my host.

ATM I’m running Proxmox with Nixos VMs running mixed docker/podman containers. It works out pretty well for my use case and with some opentofu fiddling I have most of my infrastructure defined in config files if that’s a rabbit hole you want to go down.

As for my experience with docker vs podman I have to say podman can pretty much do whatever docker does with the exception of docker swarm. You may have to do some digging to handle more advanced networking/gpu setups.

The only thing I have running docker atm is a gluetun container because container to container networking took more than 15m of research with podman so I fell back to the very well documented path using docker.

[-] maarvin@lemmy.world 2 points 8 months ago

My intent here was to be able to send requests to a set of requester chests supplied stations with a list of items needed for construction, resupply, and maybe personal long distance logistics. Although I found I can’t really arbitrarily send a train to a station via circuit signal without having all receiver stations except one be disabled.

[-] maarvin@lemmy.world 2 points 8 months ago

I have that blueprint book but, never actually looked at the core module. If I woulda been able to reverse engineer what the heck is going on there it mighta saved me some headache. Hindsight is a bear

[-] maarvin@lemmy.world 8 points 8 months ago

The easy solution to clock sync was to only have one clock. There's some funkiness required to TX a signal during the TX's clock cycle due to combinator timing but, I used that to my advantage on the RX side as it now only requires a single frame delay combinator to keep the signal alive 100% of the time.

[-] maarvin@lemmy.world 4 points 8 months ago* (last edited 8 months ago)

Woo! Same! I'm working on a bit packing solution ATM. Should be more UPS friendly and more resource friendly :D

I think I was leaning into the clock method as it doesn't have an upper/lower limit to circuit values or number of distinct circuits. Actually as I'm typing this I realized concurrent signals would have to share a 32 bit address space greatly limiting the min/max values. Two channels already limits me to 2^16-1 assuming I want a signed value.

[-] maarvin@lemmy.world 8 points 8 months ago

That would be way easier. I'm a dumb.

21
Multi channel radios! (factoriobin.com)
submitted 8 months ago by maarvin@lemmy.world to c/factorio@lemmy.world

So I had a problem. I wanted to be able to broadcast a set of circuit signals representing a list of required items for a given outpost. The problem was that the radio only supports green and red channels leading to N>3 outputs intermixing signals.

Solution: Use a repeating clock that sends a signal only if Z={{my_chosen_channel_num}}. There's some odd stuff going on combinator wise to keep signal values in memory while the Z timer is on a different channel. Also requires a single clock radio blueprint in order to keep receivers and transmitters in-sync with one another.

This is my first time fiddling with circuits more advanced than if X > Y inserter_enabled so if you find a way to make this better LMK

[-] maarvin@lemmy.world 3 points 10 months ago

np. love you bye ✌️

[-] maarvin@lemmy.world 2 points 10 months ago

I’m gonna have to save this so I can watch it after I fumble and fail setting up on Gleba without getting hints first 😅

[-] maarvin@lemmy.world 4 points 10 months ago

Dandadan. The little dance that Aira does is fun

view more: next ›

maarvin

joined 1 year ago