later on i may set up a raspberry pi to serve the function of being plugged into my UPS and broadcasting a network wide shutdown when the battery gets low instead of just plugged into my server and only shutting down my server but this solution seems kind of sucky
<sad_plan>
hm, sounds convinient. this way you dont have to fsck on everything in case of powerloss for longer periods of time
<sad_plan>
which may or may not lead to dataloss
<kris_>
i'm a little bit more confident in xfs than i should be but ive cut power to my xfs desktop randomly easily 500+ times and never even got it to fsc
<kris_>
fsck*
<kris_>
but NUT is cool
<sad_plan>
yeah Ive also forced off my laptop numerous times when I was xfs, no issues really
<kris_>
one thing i just noticed is that i dont have the qemu guest agent installed on any of my virtual machines, minor oversight
Guest85
joined
<kris_>
so i figure when my server catches a poweroff signal it just kills the VMs
<sad_plan>
hm, im considering utilizing local git remotes for rcpm instead of messing with tarballs and shit to keep track of package content and so on. would probably integrate better with oasis
<sad_plan>
I see
<kris_>
having a local git would be cool
<sad_plan>
what are you even using all those wms for?
<kris_>
i treat virtual machines like containers because i dont care much about overhead
<sad_plan>
yeah, Im having a hard time with some of this stuff in rc. and just using git instead would leviate some of those things
<sad_plan>
ah
<kris_>
on one box i have a wireguard vm and a vm for my dns proxy, that's just a little thin client that i use for network stuff separate from my big server
<sad_plan>
I would also not have to worry about tarballs at all really, because I could just remove the files in the repo. and revert if I wanna reinstall
<kris_>
big server at the moment isnt really doing much, just jellyfin and where i put files but soon i intend on hosting a few more things
<kris_>
potentially a local forgejo and some sort of backup solution
<kris_>
right now my backups are just "copy and paste onto my server over sftp when i remember to do it"
<kris_>
i still need to try oasis, is it actively maintained/developed
<sad_plan>
how many servers do you got really? when you talk about it, it kinda sounds like you run a server farm of sorts
<sad_plan>
well kinda. mcf pushes some stuff from time to time, but mostly its just me making prs, and rnpnr
<sad_plan>
mcf is really busy afaik
<kris_>
two VM hosts at the moment so nothing big
<sad_plan>
with what, I have no clue
<sad_plan>
I see
<kris_>
my network setup is just way more complicated than it needs to be more or less
<kris_>
at some point i may see about renting out some rack space in a datacenter and starting my own vps provider
<kris_>
i feel like it's a really low bar as far as that goes, all cloud providers are terrible
* kris_
wishes he lived in switzerland where you can easily get like 20 gig fiber to your house
<Guest85>
has anyone else recently been having issues compiling mesa? installed gkiss onto one of my older laptops a couple days ago and synced with repositories using kiss u and then attempting to kiss b swayfx. it can get close to finishing the compilation, but eventually stops, complaining about there being 'no matching function for call to
<Guest85>
'clang::CompilerInstance::createDiagnostics(clang::TextDiagnosticPrinter*)'. the error always happens while compiling mesa, and i haven't been using any binary repos.