<<Prev Auto Refresh Scroll to Bottom
Stuff goes here
L1[00:37:38] <Amanda> Hrm
L2[00:37:47] <Amanda> %choose iced creams or be good
L3[00:37:48] <MichiBot> Ama​nda: A wizard is never late, and sometimes engages in some "iced creams".
L4[00:39:13] <Amanda> Learning so much about wizards today
L5[00:40:05] <Amanda> But I think I'll be good, mostly because I see the dentist tomorrow
L6[01:32:44] <Amanda> Hrm
L7[01:33:58] <Amanda> Given what stephan48: said about metallb I think I'm going to have to stop serving AAAA records, because it almost feels like Verizon changes my ipv6 range more often than my ipv4 address and I can't really just nat it like I can ipv4
L8[01:38:16] <Amanda> Not to mention idk if loadbalancer objects can even have their ips changed
L9[01:44:14] <CompanionCube> %tonkout
L10[01:44:14] <MichiBot> I'm sorry Compan​ionCube, you were not able to beat Va​ur's record of 21 hours and 9 seconds this time. 5 hours, 22 minutes and 17 seconds were wasted! Missed by 15 hours, 37 minutes and 51 seconds!
L11[01:46:02] * Amanda meows at Elfi about how it's super weird that she dreamt about whatever she dreamt about last night. Something mundane but also weird
L12[01:46:17] * Amanda decided to call it early zzzmews
L13[01:46:21] <Amanda> Night girls
L14[02:02:43] <elytr​aByte> https://paste.pc-logix.com/osucohoweg
L15[02:05:12] <Hawk777> You’re not actually including the angle brackets, right?
L16[02:05:23] <Hawk777> Those aren’t part of the component address in the string.
L17[02:05:40] <elytr​aByte> no that is to represent the address of the energy cube "a52dcf94-6b4f-4e99-a6ce-a3973514b1f0"
L18[02:06:04] <Hawk777> OK. And when you say “Giving the address of the component yields a nil value” you mean you’re e.g. printing out the “proxy” variable?
L19[02:06:42] <Hawk777> Or “proxy.address”?
L20[02:06:49] <elytr​aByte> yes for now
L21[02:07:09] <elytr​aByte> I tried to do some more advanced stuff but the issue is the proxy value is nil for some reason
L22[02:07:31] <Hawk777> Hm, interesting. And the object you’re trying to control is connected to the computer case via OC cable or other devices (e.g. screens)?
L23[02:08:06] <elytr​aByte> yes I have an adapter connected to my energy cube and then cables going into the computer case
L24[02:08:31] <Hawk777> Does it show up when you use “component.list” from that library, or run the “components” program outside the Lua interpreter at the command line?
L25[02:08:33] <Mim​iru> if you type `components` at the shell does it list the cube?
L26[02:09:00] <Corded> > <Mic​hiyo> if you type `components` at the shell does it list the cube?
L27[02:09:01] <elytr​aByte> yes
L28[02:09:37] <elytr​aByte> (are the bots irc? Should I avoid images? I have some images that might elucidate things lol)
L29[02:09:55] <Hawk777> Yeah, but I think the bot turns images into URLs that we IRC users can open.
L30[02:10:10] <Mim​iru> also `component.proxy(componment.get('a52dcf94'))` will get also work, and be a lot less typoprone
L31[02:10:14] <Mim​iru> %discorded
L32[02:10:16] <MichiBot> Cor​ded is a relay between IRC and Discord. It uses a webhook (most of the time) to send IRC messages to discord, which is why there is a Bot tag
L33[02:10:26] <Mim​iru> * also work, and be a lot less typoprone
L34[02:10:38] <Corded> > <Mic​hiyo> also `component.proxy(componment.get('a52dcf94'))` will al…
L35[02:10:38] <elytr​aByte> I have just been pasting the address hehe
L36[02:10:39] <Hawk777> (Also is there something wrong with Corded? From my end on IRC I see Mimiru posting that message, and then the quoted reply appears to come from Michiyo instead)
L37[02:10:59] <elytr​aByte> I am trying to do this but no dice https://discordembeds.pc-logix.com/live/1705371058771.png
L38[02:11:04] <Michiyo> You mean the edit?
L39[02:11:10] <Michiyo> There was an edit, and then a reply
L40[02:11:31] <Mim​iru> https://discordembeds.pc-logix.com/live/1705371091516.png
L41[02:11:33] <Hawk777> Oh, was the > prefix an edit?
L42[02:11:53] <Mim​iru> So... your issue is... don't use `local` when you're working in the REPL.
L43[02:11:57] <Hawk777> Nah I was talking about the “if you type `components` at the shell does it list the cube?” line.
L44[02:12:06] <Mim​iru> local in the REPL is local to that line ONLY
L45[02:12:19] <elytr​aByte> o
L46[02:12:28] <Mim​iru> Right, I edited the post, and then I was replied to, that's the > prefix.
L47[02:12:35] <elytr​aByte> let me test!
L48[02:12:48] <Hawk777> Yeah, what I’m saying is, on my end, when elytraByte replied, it said they replied to Michiyo, not you!
L49[02:13:14] <Mim​iru> I *AM* Michiyo
L50[02:13:21] <Mim​iru> Michiyo is my account name.
L51[02:13:23] <Hawk777> Oh. OK. Sorry, never mind then.
L52[02:13:29] <Mim​iru> and Nicknames are broken for some things currently
L53[02:14:09] <Mim​iru> When pulling replies the bot is handed just the account name, and failes to pull the nickname for some reason ._.
L54[02:14:15] <Mim​iru> so yes, THAT part is broken
L55[02:17:13] <elytr​aByte> Thank you! it works as anticipated now. https://discordembeds.pc-logix.com/live/1705371433297.png
L56[02:17:21] <Corded> > <Mic​hiyo> I *AM* Michiyo
L57[02:17:21] <Klo​nade> *gasp* plot twist
L58[03:29:52] <Mim​iru> :P
L59[03:42:25] ⇦ Quits: Hawk777 (~Hawk777@2001:569:7e40:1300:2edc:6cb6:c5b:b9f5) (Quit: Leaving.)
L60[03:47:59] ⇨ Joins: Hawk777 (~Hawk777@2001:569:7e40:1300:39f2:bb78:7d2b:da5b)
L61[04:59:52] <Forec​aster> %sip
L62[04:59:53] <MichiBot> You drink a wonderful copper potion (New!). Forecaster's skin turn the color of bavarium but with a chocolate glow until hell has a mild winter.
L63[05:00:08] <Forec​aster> Wut
L64[08:24:23] ⇨ Joins: Mariflat27 (~Mariflat2@46.98.129.184)
L65[08:24:50] ⇦ Quits: Mariflat27 (~Mariflat2@46.98.129.184) (Client Quit)
L66[08:34:21] ⇦ Quits: Hawk777 (~Hawk777@2001:569:7e40:1300:39f2:bb78:7d2b:da5b) (Quit: Leaving.)
L67[10:02:08] <stephan48> Amanda: they can
L68[10:03:23] <stephan48> Amanda: what you can do is have a cron checking your v6 range and then update it & restart the metallb controller
L69[10:04:44] <stephan48> Also there is external-dns if you have not stumpled over this already which allows updating external dns providers dynamically with cluster internal(f.e. LoadBalancer) records
L70[10:09:40] <Izzy> guess who made questionable financial decisions today
L71[10:09:44] <Izzy> <-- this individual
L72[10:09:46] <Izzy> https://shadowkat.net/tmp/2c38.html
L73[10:18:29] <Forec​aster> hrm
L74[10:19:05] <Forec​aster> to move sorting duties to the script on this computer I need to create an API to query available space on the media server...
L75[10:19:15] <Forec​aster> because that's hard to do over FTP I think
L76[10:20:15] <Forec​aster> adding an API for that to the existing interface isn't too difficult really, but it's still going to require some re-organizing
L77[11:24:30] ⇨ Joins: Mariflat (~Mariflat@46.98.129.184)
L78[11:25:35] ⇦ Quits: Mariflat (~Mariflat@46.98.129.184) (Remote host closed the connection)
L79[12:17:50] <Amanda> stephan48: problem is I don't think anything is _told_ that the range changed. I only know when stuff starts taking forever over v6 and then I renew the lease on my servers
L80[12:18:12] <Amanda> Unless I can grab that info out or unifi somehow
L81[12:21:37] <Amanda> Rather, when stuff starts timing out over v6
L82[12:40:07] <Amanda> Rather, the /128 never changes, not sure about the /64 or whatever gets assigned
L83[13:02:44] <Amanda> %octime
L84[13:02:44] <MichiBot> Ama​nda: 2024-Jan-16 13:02:44
L85[13:15:35] ⇦ Quits: Corded (~Corded@2001:19f0:8001:d8f:5400:4ff:fea5:61b0) (Remote host closed the connection)
L86[13:15:38] ⇨ Joins: Corded (~Corded@2001:19f0:8001:d8f:5400:4ff:fea5:61b0)
L87[13:15:38] zsh sets mode: +v on Corded
L88[13:28:22] <Amanda> stephan48: so, I might not have a choice anyway. It seems only private-range ipv6 is being assigned to my nodes, as well as a /128, which kubectl isn't picking up
L89[13:28:47] <Amanda> I'm guessing Verizon isn't giving me a large enough range for all my servers
L90[13:28:51] <Amanda> * devices
L91[13:29:20] <stephan48> Did you try prefix delegation?
L92[13:29:28] <Amanda> That's what I'm using IIRC
L93[13:29:31] <stephan48> Via dhcp v6
L94[13:30:29] <Amanda> wait
L95[13:31:50] <stephan48> Try requesting /56 or /59 on your router - then it should allow you to get a /64 via dhcp v6 on your server from your router
L96[13:33:09] <Amanda> Do I need to do anything special on the debian side for this? My interfaces contains this:
L97[13:33:09] <Amanda> ... long message truncated: https://matrix.camnet.site/_matrix/media/v3/download/camnet.site/wCvNhBNzyvVmwHPFvwkHNPUJ (4 lines)
L98[14:20:15] <Amanda> %remindme 1h48m check if regina-prime has a /64 now
L99[14:20:15] <MichiBot> I'll tell you "check if regina-prime has a /64 now" in 1h48m at 01/16/2024 04:08:15 PM
L100[14:23:39] <Amanda> Apparently Debian required additional configuration to tell it to also request a range with PD
L101[14:33:58] <Forec​aster> %sip
L102[14:33:59] <MichiBot> You drink a boiling aqua potion (New!). After drinking the potion Forecaster realizes the bottle has their face on it.
L103[14:37:56] <Amanda> stephan48: question, could my requesting a /60 Prefix-delegation from verizon when apparently they support a /56 cause my v6 range to seem to change more often?
L104[14:40:23] <stephan48> Dont think so
L105[14:40:30] <Amanda> damn
L106[14:40:34] <stephan48> :(
L107[14:40:34] <Amanda> was hoping it might be the case
L108[14:41:00] <stephan48> But a running dhclient should automatically tell you about the new prefiy
L109[14:41:12] <stephan48> Prefix so you can adapt it
L110[15:00:35] <Amanda> TIL kvm64 doesn't include x86_64_v3 instructions. This explains why my CI for ansible suddenly started failing when mise started allowing precompiled python
L111[15:38:45] ⇨ Joins: Vexatos (~Vexatos@p200300eaef04fd00206defafd891d06f.dip0.t-ipconnect.de)
L112[15:38:45] zsh sets mode: +v on Vexatos
L113[15:45:57] ⇨ Joins: Maxze10 (~Maxze10@176.59.206.71)
L114[15:46:36] ⇦ Quits: Maxze10 (~Maxze10@176.59.206.71) (Client Quit)
L115[16:08:16] <MichiBot> Amanda REMINDER: check if regina-prime has a /64 now
L116[16:24:34] <Amanda> nope.
L117[16:45:11] <Forec​aster> The nerve!
L118[16:48:05] <Amanda> I agree, it's very rude of it
L119[17:25:06] ⇨ Joins: Maxze10 (~Maxze10@176.59.206.71)
L120[17:25:51] ⇦ Quits: Maxze10 (~Maxze10@176.59.206.71) (Client Quit)
L121[17:37:27] ⇨ Joins: Maxze10 (~Maxze10@176.59.206.71)
L122[17:37:37] ⇦ Quits: Maxze10 (~Maxze10@176.59.206.71) (Client Quit)
L123[18:11:27] ⇦ Quits: Amanda (~m-yt727s@pool-108-16-20-147.phlapa.fios.verizon.net) (Read error: Connection reset by peer)
L124[18:13:20] ⇨ Joins: Amanda (~m-yt727s@pool-108-16-20-147.phlapa.fios.verizon.net)
L125[18:42:54] <stephan48> sadly the Deutsche Terrorkom aka my provider keeps v4 and v6 stable aslong as the upstream connection does not drop
L126[18:43:15] <stephan48> might play with it over the weekend to see how propagation of the new prefix really works for ipv6
L127[18:43:40] <Amanda> Okay, apparently I didn't manage to enable ipv6
L128[18:44:15] <Amanda> Dispite rebuilding the cluster several times
L129[18:46:46] <stephan48> before you hunt adinfintatem..... delete the node in k3s and and then restart the cluster
L130[18:47:16] <stephan48> k3s is really fucking annoying in that regard, also bitches sometimes if you don't explicitly configure an ipv6 node ip
L131[18:47:23] <Forec​aster> maybe you need to rebuild it some more
L132[18:47:36] <Forec​aster> then you can say "Despite rebuilding it several *more* times"
L133[18:48:35] <Amanda> stephan48: discovered something fun, my ceph rbd isn't fast enough for etcd apparently
L134[18:48:47] * stephan48 goes to die internally
L135[18:48:49] <stephan48> yes
L136[18:49:08] <stephan48> thats why i am paying hetzner for 2x2TB nvmes right now
L137[18:49:09] <Amanda> And I already ran k3s-unsintall.sh again. :D
L138[18:49:21] <stephan48> my rotating disks where too slow for k8s
L139[18:49:47] <Amanda> How did you install it anyway? I'm trying to use an ansible role
L140[18:50:07] <Amanda> ( after I already did it manually once )
L141[18:50:45] <stephan48> rootless per documentation and then manually messed with a wrapper i wrote to magically set network up in a totally crazy way
L142[18:50:57] <Amanda> ah
L143[18:51:22] <stephan48> and on my root i used talos os which is a k8s distribution for "oh i have not had any pain this decade", it is basically just an API with k8s running on it
L144[18:51:48] <stephan48> you can only modify the OS via configuration uploads which essentially rebuild the OS under the node... like a locked down nixos
L145[18:52:07] <stephan48> i plan to opensauce my documentation when its in a presentable state
L146[18:52:21] <Amanda> ah
L147[18:53:05] <stephan48> but essentially my goal is to run k3s rootless with multiple nodes but connect it to my network via DHCP and announcing internal ranges via BGP... which i also use for metallb so Loadbalancing works on ULA and global addresses.
L148[18:53:16] <stephan48> my friend just said: you are crazy and walked away
L149[18:53:32] <stephan48> it sort of works but still needs some cleanup
L150[18:53:39] <stephan48> and is 99% rootless.
L151[18:53:49] <Mim​iru> "Self why don't you bother to do more advanced stuff like this?", reads the above "OOOOOOH, right... that's why"
L152[18:54:24] <Amanda> Join us, @Mimiru, the insanity mines are warm this time of year~
L153[18:54:25] <Forec​aster> Advanced things bring advanced problems
L154[18:54:35] <stephan48> "self but i still have sanity self" "yes and only because you do not bother with that shit"
L155[18:55:02] <stephan48> @Mimiru, but that is just the networking part? :P
L156[18:55:09] <Mim​iru> Amanda, Nah, I'm good. Thanks though.
L157[18:56:02] <Mim​iru> I'm perfectly fine over here with my Proxmox and Hyper-V, and VERY VERY (Did I mention VERY) limited use of Docker.
L158[18:56:36] <stephan48> fine and Hyper-V in one sentence? common.
L159[18:58:43] <Liizzii> lol, our hyperv clusters at work (3 nodes in each) are fine if you pause and drain a node before shutting it down but if a node were to suddenly up and die, the remaining nodes go into a weird "AAAAAAAAAAAAAA" state and the whole thing is sluggish as fuck until either some timeout elapsed and the remaining nodes consider the lost node completely
L160[18:58:43] <Liizzii> dead or the lost node comes back into the cluster
L161[18:59:06] <Michiyo> I really wish I could do HA for my Hyper-V setup... :(
L162[18:59:13] <Forec​aster> sounds fun
L163[18:59:57] <Forec​aster> for a certain definition of fun
L164[19:00:56] <Liizzii> on the plus side though, today I finally managed to move the 3rd node on one of our clusters to a different rack and have it's 4x1G LAG port(s) moved to the same switch as the other two so now me having a momentary blip in the connection between the old switch and new core switch wont throw that cluster into a fit
L165[19:05:39] <Mim​iru> The issue I have with HA on my Hyper-V stuff is... no clustered filesystem :/ No way for them to share storage and one of these VMs is ~500GB so not easy to just... move around if I need ot shut down a node.
L166[19:06:04] <Mim​iru> Really wish Hyper-V could access vhdxs via SMB.. lol
L167[19:06:32] <Forec​aster> I nearly shut down my computer...
L168[19:06:46] <Forec​aster> still have 110 voyager episodes to go
L169[19:06:49] <Mim​iru> whoops
L170[19:06:53] <Mim​iru> Good save. lol
L171[19:07:13] <Liizzii> our clusters go via 2x10Gig iSCSI networks to a Dell compellent storage box, though the volume setup on it seems really stupid
L172[19:07:48] <Mim​iru> I did iSCSI, but I couldn't connect multiple clients to it at once.
L173[19:08:03] <Amanda> Maybe I should have three controller nodes for k3s like I did with nomad, all running on the pve hosts
L174[19:08:11] <Mim​iru> I suppose I could have just been REALLY careful to only connect one client at a time.
L175[19:27:15] <Amanda> grr
L176[19:27:17] <Amanda> * Service "traefik" is invalid: spec.ipFamilies[1]: Invalid value: "IPv6": not configured on this cluster
L177[19:29:03] <Amanda> Hrm, is it because the controller nodes don't have ipv4 InternalIP entries
L178[19:29:09] <Amanda> er, ipv6
L179[19:30:35] <stephan48> did you configure v6 pod and service ranges?
L180[19:31:38] <Amanda> yes, but apparently the ansible role didnt copy them. :D
L181[19:32:16] <Amanda> https://matrix.camnet.site/_matrix/media/v3/download/camnet.site/zQAaUaKTAawQFQckHYpOTyzj/Screenshot%20from%202024-01-16%2014-32-06.png
L182[19:32:28] <Amanda> Which is unfortunate, because they need to be there on cluster init. :D
L183[19:32:29] <stephan48> whoopsi
L184[19:32:32] <stephan48> yes
L185[19:32:53] <stephan48> or rather "there needs to be a valid v6 value on cluster init"
L186[19:33:15] <stephan48> technically nothing breaks if that changes after first start, *cough cough* ofcourse it should never change
L187[19:53:43] <Forec​aster> %sip
L188[19:53:44] <MichiBot> You drink a soft spice potion (New!). A genie tries to turn Forecaster into a dryad, Forecaster tries to evade but is caught in the beam and transformed until they see a bird. (4 vs DC 12)
L189[19:54:11] <Forec​aster> 😐
L190[19:54:41] <Forec​aster> Good thing there's plenty of birds in the forest
L191[19:57:03] <Amanda> I'll try again with all the agent/server vars in a single group's vars it feels like this ansible role is a bit loosy-gooey with where it pulls variables from
L192[19:57:05] <Amanda> Unfortunately atm I need to see the toof doctor
L193[19:58:19] <stephan48> good luck! always remember, fighting back is an option
L194[20:00:39] <Forec​aster> Use your teeth
L195[20:01:00] <Forec​aster> What are they gonna do, tell you not to?
<<Prev Auto Refresh Scroll to Top