2023-08-02 11:47:30 <clandmeter> our cdn is doing roughly 2.5PB per month. that's kind of crazy.
2023-08-02 11:50:30 <mps> what is SPB
2023-08-02 11:57:37 <clandmeter> 2.5 PB
2023-08-02 11:57:48 <clandmeter> there is no S :)
2023-08-02 11:58:48 <mps> oh, I'm blind or have to increase brightness :)
2023-08-02 12:00:12 <mps> iiuc, alpine download is 2.5 petabytes monthly
2023-08-02 12:00:52 <clandmeter> yes
2023-08-02 12:01:56 <mps> what that means? a lot or a little
2023-08-02 12:02:26 <clandmeter> depends how you look at it i guess
2023-08-02 12:02:42 <clandmeter> im not sure what other distros do
2023-08-02 12:04:33 <mps> hm, maybe this is not important (and it is not important to me)
2023-08-02 12:04:53 <minimal> obviousl that excludes all the docker pulls of Alpine, which I'm sure would be large
2023-08-02 12:07:28 <clandmeter> mps: it says something about the popularity of the project
2023-08-02 12:07:35 <clandmeter> if that concerns you or not, that i dont know.
2023-08-02 12:08:33 <clandmeter> minimal: yeah docker pulls not, but i guess what happens in those containers could influence it.
2023-08-02 12:08:55 <mps> how could that says something about popularity? maybe I have cron running somewhere to download all pkgs every minute ;)
2023-08-02 12:09:19 <clandmeter> haha, i dont think you will see that in this scale :)
2023-08-02 14:16:33 <ncopa> wow... 
2023-08-02 14:16:38 <ncopa> thats insane
2023-08-02 14:19:12 <ncopa> alpine had 9.5M pulls last week
2023-08-02 14:19:27 <ncopa> 9.4 to be more exact
2023-08-02 14:20:25 <ncopa> ubuntu had 29M
2023-08-02 14:20:44 <ncopa> ubuntu has approx 3x pulls compared to alpine
2023-08-02 15:33:11 <ikke> funny, the dendrite test suite on s390x is failing with the same error that we have in our environment
2023-08-02 17:05:25 <ncopa> https://gitlab.alpinelinux.org/alpine/aports
2023-08-02 17:05:38 <ncopa> Bad Gateway
2023-08-02 17:05:45 <ncopa> fatal: unable to access 'https://gitlab.alpinelinux.org/alpine/aports.git/': The requested URL returned error: 502
2023-08-02 17:05:55 <ncopa> ikke:^^^
2023-08-02 17:05:58 <ikke> ncopa: needed to restart gitlab
2023-08-02 17:06:04 <ikke> for security upgrades
2023-08-02 17:06:07 <ikke> will be back in a bit
2023-08-02 17:06:08 <ncopa> ok :)
2023-08-02 17:06:16 <ncopa> got scared for a sec there
2023-08-02 17:07:06 <ikke> it's back
2023-08-02 17:07:16 <ncopa> thank you sir!
2023-08-03 16:34:41 <ptrc> ikke: rv64 is stuck again
2023-08-03 16:35:24 <ptrc> been stuck for quite a while actually
2023-08-03 16:35:34 <ikke> kicked the process
2023-08-03 16:35:44 <ptrc> maybe there should be a "last update" timestamp on buol
2023-08-03 16:35:48 <ptrc> build*
2023-08-03 16:35:58 <ikke> ptrc: was thinking the same, didn't get to it adding it
2023-08-05 13:50:00 <mps> is the build-edge-x86_64 stuck
2023-08-05 13:50:37 <ikke> No, seems active
2023-08-05 13:51:47 <mps> hm, long time build electron
2023-08-05 13:52:08 <ikke> It's huge
2023-08-05 17:20:54 <mps> algitbot: retry master
2023-08-06 07:04:56 <ikke> gitlab received a lot of requests last night
2023-08-09 06:45:48 <clandmeter> ikke: around?
2023-08-09 07:04:15 <clandmeter> ikke: do you have an idea why this would fail? https://gitlab.alpinelinux.org/alpine/aports/-/pipelines/173626
2023-08-09 07:04:31 <clandmeter> it suddenly complains about permissions
2023-08-09 07:04:40 <clandmeter> which it did not before
2023-08-09 07:08:36 <clandmeter> if there are no permission changes, i guess it is related to some gitlab version update
2023-08-09 09:00:39 <clandmeter> ncopa: we got an email from fastly, but i think we are not affected. i guess you didnt see it yet.
2023-08-09 09:00:55 <ncopa> i saw it
2023-08-09 09:01:10 <ncopa> and i also thought we are not affected, since we dont use azure?
2023-08-09 09:01:26 <ncopa> reminds me, we should work on alpine azure cloud images
2023-08-09 09:02:52 <ncopa> clandmeter: another thing I though. the aws aarch64 cloud images only works with virt, as I undrestand. Do you think that https://gitlab.alpinelinux.org/clandmeter/alpine-disk-image could work on amazon's arm  bare metal machines?
2023-08-09 09:04:31 <ikke> ncopa: yes, we should certainly work on azure images 
2023-08-09 09:04:43 <clandmeter> could be, im not sure as i never used amazon metal stuff
2023-08-09 09:05:08 <ncopa> i suppose we only need use the amazon driver for tinycloud
2023-08-09 09:05:41 <ncopa> what would be nice was if we had generic cloud image, which would autodetect aws, equinix or azure
2023-08-09 09:05:57 <ncopa> or alpine-config/NoCloud
2023-08-09 09:07:46 <clandmeter> a generic image would be nice, but it would mean we need to ship it with 2 kernels.
2023-08-09 09:07:49 <clandmeter> or have 2 different iamges
2023-08-09 09:08:51 <clandmeter> equinix also does some magic on the images before pushing it
2023-08-09 09:18:19 <ncopa> im thinking two different images. virt and bare-metal
2023-08-09 09:18:36 <ncopa> still better than one image per cloud?
2023-08-09 09:19:42 <clandmeter> yes, that for sure.
2023-08-09 09:21:37 <clandmeter> did you already succeed to build it without root and with regular tools instead of packer?
2023-08-09 09:22:01 <ncopa> sort of yes
2023-08-09 09:22:12 <ncopa> we do that in mirantis
2023-08-09 09:22:38 <ncopa> we have a docker -> disk image script
2023-08-09 09:23:10 <ikke> clandmeter: did I see the pipeline worked the 2nd time?
2023-08-09 09:23:27 <clandmeter> could that translate into our build scripts?
2023-08-09 09:23:38 <ncopa> possibly
2023-08-09 09:23:42 <clandmeter> i guess we do not want to introduce docker in our build process 
2023-08-09 09:24:08 <ikke> clandmeter: I also should respond to IBM about the 2nd ppc64le machine, didn't get to it yet 
2023-08-09 09:24:09 <clandmeter> ikke: yes i fixed it by adding ncopa as developer in my project
2023-08-09 09:24:26 <clandmeter> its weird as ncopa has admin privs 
2023-08-09 09:24:44 <ikke> clandmeter: ah, sounds like a ci_job_token permission change 
2023-08-09 09:25:11 <clandmeter> i have been thinking of just movign it to the alpine namespace
2023-08-09 09:25:45 <clandmeter> i guess the permissions would just work if in /alpine
2023-08-09 09:26:29 <clandmeter> ncopa: btw, i do not include tinycloud in my image. its very bare..
2023-08-09 09:26:48 <clandmeter> but of course it can be added if needed
2023-08-09 09:34:41 <ncopa> ok
2023-08-09 10:58:42 <mps> I could take maintainership of main/bcache-tools, currently pkg is outdated and is renamed upstream to bcachefs-tools. any objection?
2023-08-09 10:59:09 <ikke> Not from me
2023-08-09 10:59:34 <ikke> mps: don't forget to add provides and replaces if you rename it
2023-08-09 10:59:46 <mps> ikke: sure
2023-08-09 11:01:20 <mps> question is, should the new pkg go in testing or directly to main? I think main but ask just in case I overlooked something
2023-08-09 14:54:52 <rootwyrm> So, hopefully quick(ish) question.. looking at u-boot-libretech specifically. But it's not clear to me exactly what type of u-boot build it is.
2023-08-09 14:55:20 <ikke> mps: ^
2023-08-09 15:04:50 <mps> ikke: I don't understand question? what aport?
2023-08-09 15:08:05 <mps> rootwyrm: could you be more precise with question
2023-08-09 15:12:55 <rootwyrm> mps: what format/part of u-boot is it building? Just the non-EFI image and not the boot-block image?
2023-08-09 15:14:19 <mps> rootwyrm: I don't have this board so I didn't tested. but in new u-boot EFI should be enabled
2023-08-09 15:14:52 <mps> s/enabled/built in/
2023-08-09 15:20:11 <rootwyrm> mps: well, good news and bad news then... I have two of the aml-s905x-cc (v1's. The v2's different.) So hypothetically it should be 'dd if=the_uboot_bin of=sdcard seek=1' buuuuut, the s905x can't find it.
2023-08-09 15:20:45 <rootwyrm> (And let me tell you, I have _THOUGHTS_ about the libre team. "JUST COPY SOMEBODY ELSE'S KERNEL IN", really?)
2023-08-09 15:22:00 <mps> there are too much SBCs around and I don't know how to setup much of them.
2023-08-09 15:23:07 <mps> if you want to run alpine on them you can open bug report and explain what is have to be done with u-boot, kernel etc...
2023-08-09 15:23:07 <rootwyrm> Well, it definitely doesn't help that AMLogic has a bunch of 'secure firmware' crap and hides everything useful behind customer-only NDA.
2023-08-09 15:23:54 <mps> we only use arm-trusted-firmware where it is needed and is possible
2023-08-09 15:25:16 <rootwyrm> So, it's probably missing, because the s905x setup is deeply, deeply stupid. It specifically will only look for the bootloader in a specific format and basically implodes if you dare try proper UEFI and GPT.
2023-08-09 15:26:49 <mps> understand your feelings
2023-08-09 15:27:36 <rootwyrm> Yeah.. you can sorta-chainload UEFI, hypothetically. But there's some AMLogic proprietary blobs that are probably not in there :(
2023-08-09 15:28:02 <rootwyrm> And of course, you need AMLogic's written permission to distribute any part of them.
2023-08-09 15:28:22 <mps> libretec is added by commit bc40aac7c79ddf1c13184678666eed14a2b548b0
2023-08-09 15:29:42 <rootwyrm> mps: yeah, and I don't see any of the extra amlogic patches. But it's also clear as mud if u-boot actually got to mainline any of them.
2023-08-09 15:30:34 <mps> well, alpine will not add non free things
2023-08-09 15:31:00 <mps> and u-boot upstream, I think
2023-08-09 15:31:06 <rootwyrm> Yeah, and couldn't anyways, because nobody can distribute the blobs without amlogic's permission.
2023-08-09 15:33:17 <rootwyrm> Also why my response to them has been "just copy a non-free kernel, are you people stupid or just lying about 'open' like everyone else?"
2023-08-09 15:35:09 <minimal> rootwyrm: isn't that just the usual Chinese SBC working practices, everyone distributing "random" binaries that they don't know where they came from and no-one has the (any or complete) source for
2023-08-09 15:36:47 <rootwyrm> minimal: usually they don't even bother to distribute any binaries, so, technically a big step up here!
2023-08-09 15:41:12 <minimal> I'm thinkinf of BSPs that are distributed between CPU designers, board manufacturers, shipping product companies etc
2023-08-09 15:42:16 <mps> 
ACTION don't buy too much closed hardware

2023-08-09 15:47:59 <rootwyrm> ah-ha, found the commit. https://github.com/u-boot/u-boot/commit/8bbfb40e572ddb8a11482ca513131e21a75e99de
2023-08-11 13:28:14 <raspbeguy> Hello, I was wondering if any help was needed on the infrastructure maintainance. If yes, would be glad to help.
2023-08-11 13:28:40 <PureTryOut> Is Gitlab slow for anyone else currently or is it just me?
2023-08-11 13:30:02 <PureTryOut> Oh after being unable to load anything for like an hour, now I send the previous message everything loads again lol. So nvm I suppose
2023-08-11 14:30:03 <ncopa> raspbeguy: appreciate that! clandmeter, ikke: ^^^
2023-08-11 14:30:38 <ikke> Yes, I've directed raspbeguy here :⁠-⁠)
2023-08-11 14:33:23 <rootwyrm> So I finally figured out what's up with aml-s905x-cc; it's the BL31 crap. There is an open source option for the GXL encryption that doesn't need any of the blobs, but need to do more testing and fussing around.
2023-08-11 14:35:16 <mps> ikke: I see on my lxc this process '14838 pts/6    Z      7:50 [firefox] <defunct>'
2023-08-11 14:35:48 <mps> is there a way to kill it, or the only is to restart lxc
2023-08-11 14:36:09 <ikke> What process owns it?
2023-08-11 14:37:29 <mps> no one, it is zombie
2023-08-11 14:37:46 <mps> oh wrong
2023-08-11 14:38:29 <mps> ikke: sorry for noise, I killed it
2023-08-11 14:39:51 <ikke> Np
2023-08-11 14:40:16 <ikke> A zombie process means it's parent didn't check on yet 
2023-08-11 14:40:30 <mps> yes yes
2023-08-11 14:40:48 <mps> but I missed proper `ps`
2023-08-11 14:41:22 <mps> ncopa: should we merge !49909 to 3.18
2023-08-11 14:41:44 <rootwyrm> mps: generally speaking, how hard would it be to have the u-boot-libre remain a meta but have a fairly significant set of build changes?
2023-08-11 14:42:14 <ncopa> mps: i'd say no, unless there is a strong reason to do so. like a security vulnerability
2023-08-11 14:43:21 <mps> ncopa: long bugfix list make me to fear
2023-08-11 14:44:00 <mps> 'All: fix a number of memory leaks and unreachable/harmless integer overflows
2023-08-11 14:44:53 <mps> rootwyrm: only if there are free patches
2023-08-11 14:46:04 <ncopa> mps: I'll have a look at it next week
2023-08-11 14:46:45 <mps> ncopa: ok, I will build it for my 3.18 systems locally
2023-08-11 14:48:50 <rootwyrm> mps: yeah, it's actual open source, but it necessitates some pretty big build changes. Also makes cross-build basically impossible; has to be native aarch64.
2023-08-11 14:50:08 <mps> rootwyrm: well, we build all pkgs native anyway
2023-08-11 18:08:21 <raspbeguy> ncopa, ikke: so what's the onboarding process?
2023-08-11 19:06:58 <ikke> raspbeguy: I'm away at the moment, we'll talk about it later 
2023-08-11 19:13:02 <raspbeguy> Okie dokie
2023-08-11 19:53:15 <mps> now is the time to upgrade opensbi and rebuild u-boot with it
2023-08-11 19:53:32 <mps> uhm, wrong channel
2023-08-13 15:13:57 <ikke> raspbeguy: Could you tell what kind of experience you have with infrastructure / system administration? I'll see if there is something you can help with
2023-08-13 15:17:55 <raspbeguy> ikke: I work as SRE at OVH. Working with Linux machines everyday
2023-08-13 15:18:19 <ikke> What kind of technologies / tools do you have exprience with?
2023-08-13 15:18:35 <ikke> We mostly use docker and lxc
2023-08-13 15:19:13 <ikke> The infrastructure is mostly handled by clandmeter and me
2023-08-13 15:19:33 <raspbeguy> I essentially have experience with KVM and Xen but I also use docker and LXC too
2023-08-13 15:20:08 <ikke> any kind of orchestration or configuration management?
2023-08-13 15:20:13 <raspbeguy> My personnal preference is KVM using libvirt
2023-08-13 15:20:20 <raspbeguy> Ansible mostly
2023-08-13 15:20:52 <raspbeguy> Used to do some Salt and Puppet but very occasionally 
2023-08-13 15:22:03 <raspbeguy> I also did some terraform
2023-08-13 15:22:16 <ikke> Any programming experience?
2023-08-13 15:23:04 <raspbeguy> Little, some python and trying to learn Golang
2023-08-13 15:23:53 <raspbeguy> I did some C many years ago
2023-08-13 15:24:13 <raspbeguy> But didn't practice since then
2023-08-13 15:25:20 <ikke> The projects that we maintain are mostly located in https://gitlab.alpinelinux.org/alpine/infra
2023-08-13 15:25:32 <ikke> docker images: https://gitlab.alpinelinux.org/alpine/infra/docker
2023-08-13 15:29:37 <raspbeguy> If that's ok I'll continue talking tomorrow. Family reunion right now
2023-08-13 15:30:33 <ikke> raspbeguy: yeah, no worry
2023-08-13 15:30:52 <ikke> I'll discuss it with clandmeter and see where you can help
2023-08-13 18:27:07 <iandi> ikke: I would like to help (and learn new stuff on the way) as well. I'm working as sysadmin in software development company (~70 ppl). I have experience with ansible, terraform, prometheus stack for monitoring/alerting, k8s, a bit of docker. Programming experience is limited but I know bash, python and a bit of go.
2023-08-13 18:28:35 <iandi> please let me know if I could be helpful
2023-08-13 18:35:14 <ikke> iandi: thanks
2023-08-14 07:12:41 <clandmeter> good morning
2023-08-14 07:19:53 <clandmeter> ikke: that sounds great
2023-08-14 07:20:41 <clandmeter> I think we would want to arrange a conf call if that is ok with everyone.
2023-08-14 07:21:01 <clandmeter> just to have a talk about what and how we do things.
2023-08-14 08:20:15 <raspbeguy> clandmeter, why not, but please be mindful my oral english isn't that great
2023-08-14 08:44:21 <clandmeter> raspbeguy: join the club :)
2023-08-14 08:45:54 <mps> english per se "isn't that great" ;p
2023-08-14 09:03:35 <raspbeguy> ut at least most people can understand it somewhat 🙂
2023-08-14 09:06:07 <mps> imo, it is better to not understand at all than to understand superficially or wrongly
2023-08-14 11:38:45 <raspbeguy> also we could all talk Esperanto 🙂
2023-08-14 11:41:36 <mps> oh no, esperanto is artificial lang and as such it is worse than english
2023-08-14 11:57:37 <rootwyrm> English is the language that drags other languages into dark alleys, knocks them out, and rummages through their pockets for loose grammar.
2023-08-14 12:00:02 <mps> :)
2023-08-15 10:06:48 <ncopa> i just discovered NATS (https://nats.io) I think this does most what we need for building a new build infra
2023-08-15 10:07:41 <ncopa> its like MQTT + key/value store + object store + streams
2023-08-15 10:21:32 <ncopa> oh, it even has support for MQTT
2023-08-15 17:11:16 <clandmeter> seems to have enough features... now we need to wait until they switch license :)
2023-08-15 19:35:24 <ptrc> has anyone contacted the person from #13269?
2023-08-15 19:35:33 <ptrc> > May I can offer some riscv64 hardware resources like Unmatched remotely or other riscv64 hardware.
2023-08-15 19:39:41 <ikke> ptrc: nope
2023-08-15 19:40:25 <ikke> Is an unmatched board powerfull enough?
2023-08-15 19:41:50 <ikke> 5-core, 16G
2023-08-15 19:46:12 <ptrc> 5 cores sounds like definitely not enough for a builder
2023-08-15 19:47:40 <ikke> yeah, the only thing that sounded suitable was that milk board
2023-08-15 19:48:15 <ikke> milkv
2023-08-15 19:49:36 <ikke> https://milkv.io/pioneer
2023-08-15 19:55:40 <clandmeter> raspbeguy: do you have time next week for a call?
2023-08-15 19:56:07 <clandmeter> iandi: what about you?
2023-08-15 19:56:24 <clandmeter> we are cest 
2023-08-15 19:56:51 <clandmeter> France probably also
2023-08-15 20:19:56 <mps> ptrc: tell this person we have #alpine-riscv64 channel where we discuss most about riscv
2023-08-15 20:21:09 <mps> and yes, 5 core 16GB ram is not enough, lxc+qemu-user on x86_64 is a lot faster
2023-08-15 20:21:35 <mps> 'a lot' => substantially, better to say
2023-08-15 20:22:10 <mps> ikke: yes, milkv sounds good
2023-08-15 20:29:31 <raspbeguy> clandmeter: sure
2023-08-15 21:53:16 <rootwyrm> why can't arm64 have something even halfway reasonable? Sigh.
2023-08-16 09:14:37 <iandi> clandmeter: yes and I'm cest too
2023-08-16 13:23:51 <clandmeter> Can you pm me your mail addresses?
2023-08-16 13:24:19 <clandmeter> ill propose some date and times 
2023-08-16 14:31:48 <iandi> done
2023-08-16 16:21:08 <dhruvin> Is this conference call open to all? I'd like to join (to learn) if possible.
2023-08-16 16:39:16 <clandmeter> dhruvin: let me discuss with the others
2023-08-16 21:05:30 <clandmeter> raspbeguy: can you share your email?
2023-08-16 21:20:30 <raspbeguy> clandmeter: I PMd you
2023-08-19 10:10:21 <mps> x86_64 CI can't resolve qemu-project.org https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1088775#L709
2023-08-19 15:35:29 <ikke> mps: changed resolvers, works now
2023-08-19 16:20:05 <mps> ikke: I see, thanks
2023-08-21 17:54:24 <mps> anyone know or have openrc script example to run normal program in background at boot
2023-08-21 17:55:02 <mps> add '&' at the end of invocation?
2023-08-21 19:06:30 <ikke> mps: yes, you can do that in .xinitrc
2023-08-21 19:07:46 <mps> ikke: I know this, but I want in on boot from /etc/init.d/ openrc init script
2023-08-21 19:08:26 <ikke> is it a graphical application?
2023-08-21 19:08:51 <mps> no, program which control touchbar on apple m1
2023-08-21 19:09:14 <mps> https://github.com/whatamisupposedtoputhere/tiny-dfr
2023-08-21 19:09:30 <minimal> mps: /etc/init.d/local ?
2023-08-21 19:09:45 <mps> I prepared aport for alpine and want to add -openrc subpkd
2023-08-21 19:10:02 <mps> minimal: doesn't make sense for system service
2023-08-21 19:10:26 <ikke> just run it as a normal service then
2023-08-21 19:10:34 <minimal> you said you wanted to run "normal" program
2023-08-21 19:10:39 <ikke> using & only makes sense in a shell
2023-08-21 19:10:53 <mps> ikke: tried but it then blocks boot
2023-08-21 19:11:28 <mps> minimal: yes, it is 'normal' i.e. doesn't have option to run in background as daemon
2023-08-21 19:11:39 <ikke> https://github.com/OpenRC/openrc/blob/master/service-script-guide.md#dont-write-your-own-startstop-functions
2023-08-21 19:11:41 <minimal> so then put "command_background=yes" in init.d script?
2023-08-21 19:12:00 <mps> minimal: tried, and doesn't work
2023-08-21 19:12:23 <minimal> what happens?
2023-08-21 19:12:45 <mps> minimal: says that it doesn't create pid file
2023-08-21 19:12:57 <mps> openrc says
2023-08-21 19:13:25 <mps> I know how to run it with runit
2023-08-21 19:13:43 <mps> does openrc supervisor could be used for this
2023-08-21 19:13:58 <ikke> certainly
2023-08-21 19:14:25 <mps> ikke: ok, will look examples, maybe there is some hints
2023-08-21 19:14:42 <ikke> https://github.com/OpenRC/openrc/blob/master/supervise-daemon-guide.md
2023-08-21 19:14:46 <ikke> read both
2023-08-21 19:15:25 <mps> I think last one is enough ;)
2023-08-21 19:15:45 <mps> but will read both
2023-08-21 19:35:05 <mps> fine, got this url https://github.com/chadmed/asahi-overlay/blob/main/app-laptop/tiny-dfr/files/tiny-dfr.initd on asahi channel
2023-08-21 19:58:39 <rootwyrm> mps: use the openrc supervisor for that kind of thing. If you want I can probably throw it together quick.
2023-08-21 20:01:18 <mps> rootwyrm: I think it is not needed, script at above url works
2023-08-21 20:01:36 <mps> anyway thank you for offering help
2023-08-21 20:02:08 <mps> and if you want you can create openrc init script which we could add to pkg
2023-08-21 20:30:12 <ptrc> hm
2023-08-21 20:30:19 <ptrc> the gitlab x86_64 runner is not okay
2023-08-21 20:30:28 <ptrc> https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1090530
2023-08-21 20:31:40 <ikke> hmm, not a lot of information
2023-08-21 20:37:18 <ikke> maybe a temporary issue, I see jobs running now on that runner
2023-08-21 20:38:46 <ptrc> alright, i'll try retrying the failed jobs
2023-08-23 09:00:57 <mps> is it possible to set 'options="textrels"' in subpkg? it is but looks like it doesn't work
2023-08-23 09:21:16 <ikke> That's most likely executed in rootpkg and does not affect the shell that runs the checks
2023-08-23 09:24:29 <mps> looks like so
2023-08-23 09:24:52 <mps> so option is to disable textrels globally
2023-08-23 09:25:50 <mps> new qemu 8.1.0 have textrel in s390x firmware
2023-08-23 09:26:14 <mps> 'TEXTREL  /home/mps/aports/community/qemu/pkg/qemu-s390x-firmware/usr/share/qemu/s390-netboot.img'
2023-08-23 09:27:24 <mps> if we can't fix this in subpkg we have to add textrels to options globally?
2023-08-23 09:50:06 <ikke> Yes
2023-08-23 09:50:30 <ikke> Or find a way to fix the actual issue 
2023-08-23 09:50:36 <ikke> But might not be easy 
2023-08-23 09:54:53 <mps> ok, I will create MR and maybe someone could have idea how to fix just subpkg
2023-08-23 11:28:22 <mps> ikke: I added !50453
2023-08-23 11:29:08 <mps> if anyone have better solution for 'textrels' please do
2023-08-23 12:18:50 <rootwyrm> mps: quick q, have you worked with any of the RK5388 boards?
2023-08-23 13:29:48 <mps> rootwyrm: no, of the rockchip I worked only with rk3399
2023-08-23 13:57:15 <rootwyrm> mps: how is the 3399 for builds? that's the other hardware I'm looking at
2023-08-23 14:02:18 <mps> rootwyrm: SoC is well supported in kernel, problems could arise with some specific vendors add-ons
2023-08-23 14:02:59 <rootwyrm> that part's easy for me to solve, the big question is if it's got enough performance to be usable for building stuff.
2023-08-23 14:03:25 <mps> for example mainline kernel works fine on my rk3399 chromebook for more years
2023-08-23 14:04:07 <mps> ah this, not sure it is good choice for builders
2023-08-23 14:05:23 <mps> small programs could be built fine but big ones which requires RAM and CPU will need a time
2023-08-23 14:05:26 <rootwyrm> ugh, that's what i was afraid of. there really seems to be no options there short of the amperes which are just too expensive. i can't even find cavium thunderx2's used.
2023-08-23 14:06:04 <mps> though few years ago I've built firefox few times on it, take about 8 hours iirc
2023-08-23 14:06:44 <rootwyrm> oh, hell, if firefox is 8 hours that might actually be tolerable. a few hours turnaround is fine.
2023-08-23 14:07:03 <mps> rk35XX could be a lot better than rk3399, but I don't my experience
2023-08-23 14:07:53 <rootwyrm> yeah, the caveat with the 3588's is that they're like *twice* the cost.
2023-08-23 14:08:12 <mps> if I decide to by arm64 SBC nowadays it will be probably rk35XX
2023-08-23 14:09:01 <rootwyrm> yeah, and radxa has the rock5's. but $130US and up for the 8GB version.
2023-08-23 14:10:58 <rootwyrm> and for a builder you'd really need to get the 'model B' with the NVMe. but the rock4 with the 3399T and NVMe is only $70US
2023-08-23 14:12:14 <mps> depend what are build needs, small and not often or big ones
2023-08-23 14:12:49 <mps> for serious and big build tasks it is worth to give more money to get good machine
2023-08-23 14:14:14 <mps> maybe you could find used apple m1 machine for reasonable money
2023-08-23 15:27:33 <ikke> telmich: We have issues with our dendrite instance, so I'm not able to use the matrix channel atm. You mentioned last time you had some issues with dns on vigir23. I notice that ip address on br-lan changed from 192.168.1.254 to 192.168.1.1. Was that done as a fix? Because I see that /etc/resolv.conf points to the former address and so dns on the host itself is not working
2023-08-23 16:08:51 <telmich> <ikke> "nico🇨🇭: We have issues with..." <- That does not sound right. 192.168.1.254 should be the gateway of that network
2023-08-23 16:09:13 <ikke> Right. Both via ssh, as in the webif, 1.1 is set
2023-08-23 16:09:43 <telmich> I'm currently on low speed Internet, but it's surprining you can access it with the wrong gateway
2023-08-23 16:09:52 <ikke> It's the lan side
2023-08-23 18:55:01 <ikke> telmich: I've changed it back to .254
2023-08-23 19:00:17 <ikke> That's me
2023-08-23 19:19:12 <ikke> ok, that caused an ip conflict
2023-08-23 19:23:07 <ikke> telmich: there are 2 networks with the same subnet. One on wan and one on br-lan
2023-08-24 05:19:11 <ikke> telmich: something broke after I changed the br-lan ip (and reverted again). The servers are not responding to ipv6 anymore. 
2023-08-24 06:25:43 <ikke> It would perhaps already help if I knew what is responsible for the router advertisements for that prefix 
2023-08-24 07:27:48 <mps> arm machine is inaccessible again?
2023-08-24 07:38:35 <ncopa> arm ci is down?
2023-08-24 07:39:57 <ikke> Yes, I tried to fix something but instead it broke it 😑
2023-08-24 07:53:37 <clandmeter> ikke: did you do anything with the arm server on my side?
2023-08-24 07:55:39 <ikke> clandmeter: I have not accessed it recently 
2023-08-24 07:56:50 <clandmeter> ok ill turn it off 
2023-08-24 14:31:07 <ayakael7> Idk if this came up at some point, but this riscv64 hardware seems promising for a build server https://www.crowdsupply.com/milkv/milk-v-pioneer 
2023-08-24 14:31:18 <ikke> ayakael7: it did
2023-08-24 14:32:03 <ikke> we did try to apply for it
2023-08-24 14:32:18 <ikke> but they were looking for different kind of projects at the time
2023-08-24 14:34:02 <ayakael> Unfortunate! I hope they eventually make that hardware available to us. Any alternative hardware lining up for us?
2023-08-24 14:34:31 <ikke> I haven't seen anything else that comes close to that
2023-08-24 14:37:17 <ayakael> Indeed, the specs are impressive.
2023-08-24 14:41:47 <clandmeter> there was something like 128 core riscv cpu on my timeline last few days
2023-08-24 14:42:25 <ayakael> Also, were the s390x builder space constraints solved? The ones that caused us to drop dotnet s390x?
2023-08-24 14:42:42 <ikke> Not really
2023-08-24 14:43:44 <ayakael> Ok. I ask, as a developper requested powershell on s390x. I told him I'd look into it. dotnet8 is looking promissing for better support.
2023-08-24 14:43:54 <ayakael> And riscv64 is coming as well. 
2023-08-24 14:44:23 <ikke> ok
2023-08-24 14:46:12 <ayakael> dotnet8 will be in testing until dotnet7 goes EOL in May 24, so I'll have s390x in testing, going around the space constraint issue.
2023-08-24 14:46:43 <ayakael> May 2024*
2023-08-24 14:49:38 <rootwyrm> ayakael: for s390x? I think there's a semi-experimental for 7 available.
2023-08-24 14:51:16 <ayakael> rootwyrm: for packaging? Yeah, there was, but was dropped due to space constraints on the builder. Although I do keep builds going for s390x and ppc64le on my devel repo here: https://lab.ilot.io/ayakael/repo-apk
2023-08-24 14:52:19 <rootwyrm> ayakael: I mean upstream binaries, let me check with the s390x person
2023-08-24 14:53:50 <ayakael> Oh, for Alpine Linux? That would be great news. I've been doing the crossbuilds myself for use by dotnet7-stage0, but if there's upstream, even better.
2023-08-24 14:53:51 <rootwyrm> (also, lmk if you run into any alpine/musl specific weirdness. I *think* it's fully shaken out.)
2023-08-24 14:54:56 <ayakael> awesome,  I would hope so! It took a while for them to merge my patches for musl support on mono-flavored runtime.
2023-08-24 14:55:32 <rootwyrm> ayakael: well, mono was already pretty much in "nah, this is dead" mode soooo. ;/
2023-08-24 14:56:15 <ayakael> Yeaaah... indeed. But coreclr still doesn't support s390x and ppc64le (I doubt it ever will), so that's all we got for dotnet on those platforms unfortunately.
2023-08-24 14:57:18 <rootwyrm> Yeah, some of the 'supported' versus 'builds and runs fine' stuff gets weird for want of a better term. i.e. FreeBSD's probably never going to be officially 'supported.'
2023-08-24 14:59:08 <ayakael> Indeed, and christ on a broomstick was it a whole bunch of work to get linux-musl builds working on source-build haha, and Alpine Linux is "officially supported" haha
2023-08-24 14:59:36 <rootwyrm> <ahem> "Welcome to msbuild and arcade. Here is your handbasket."
2023-08-24 14:59:43 <ikke> s390x builder:
2023-08-24 14:59:46 <ikke> /dev/mapper/vg1-lv_var  393G  285G   91G  76% /var
2023-08-24 15:00:07 <rootwyrm> ooooooof... and you're doing qemu I presume?
2023-08-24 15:00:35 <ikke> no, this is s390x hardware provide by IBM
2023-08-24 15:00:39 <ikke> (which may be a VM)
2023-08-24 15:00:51 <ikke> or whatever kind if virtualization it has
2023-08-24 15:02:29 <rootwyrm> ah, those are usually ZFL LPARs on real hardware (or whatever the hell it's called this week. spectrum rainbow security for open source or watson for linux or some other crap.)
2023-08-24 15:02:46 <ikke> yes indeed
2023-08-24 15:03:03 <ayakael> That's enough space for building dotnet. The problems occur on release when edge and the new release match, so space requirementes effectively double.
2023-08-24 15:04:11 <ikke> we also need space for a new release
2023-08-24 15:04:13 <rootwyrm> ayakael: yeah, and _generally_ speaking, those should always match.
2023-08-24 15:09:35 <ayakael> rootwyrm: Btw, if ever you should need them. I keep these dotnet bootstraps generally up-to-date: https://lab.ilot.io/ayakael/dotnet-stage0/-/releases. I've got experimental riscv64 up right now for dotnet8-preview7.
2023-08-24 15:10:43 <ayakael> I just don't have the hardware to test them out yet
2023-08-24 15:21:23 <rootwyrm> same boat, alas. i'm still trying to find decent arm64 builder hardware that isn't $5k+ or apple
2023-08-24 21:01:17 <ikke> clandmeter: I've setup temporary ci runners on usa-bld-1 (the temp server from equinix)
2023-08-25 06:39:21 <raspbeguy> Hello, I can't find the dmvpn endpoint
2023-08-25 06:52:23 <mps> khost -t a dmvpn.alpinelinux.org
2023-08-25 07:21:34 <telmich> <ikke> "nico🇨🇭: something broke after..." <- That sounds as if the IPv6 address on the lan interface might have been flushed - I'll try to check it out today
2023-08-25 07:21:46 <mps> hm, even IPv6 doesn't work on arm machine
2023-08-25 07:21:46 <ikke> telmich: thanks
2023-08-25 07:23:04 <telmich> ikke: Looks about correct, only IPv4 on br-lan
2023-08-25 07:24:22 <ikke> telmich: do you also notice that 192.168.1.0/24 is configured for both wan and br-lan?
2023-08-25 07:24:48 <telmich> ikke: LAN: IPv4: 192.168.2.1/24
2023-08-25 07:25:35 <ikke> Oh, that's what I configured later to try to fix it 
2023-08-25 07:26:01 <ikke> It was 1.1
2023-08-25 07:26:18 <telmich> I believe you :-)
2023-08-25 07:27:29 <telmich> Re-adding 2a0a:e5c1:517:cafe::42/64 to the LAN interface
2023-08-25 07:28:48 <telmich> If you execute ping ff02::1%br-lan you'll see that there are a few devices responding to LL IPv6, so they'll re-assign itself an IPv6 address within 2a0a:e5c1:517:cafe::/64
2023-08-25 07:28:53 <telmich> (SLAAC is on)
2023-08-25 07:29:58 <telmich> afair, openwrt solves the 192.168.1.0/24 on both interfaces situation somehow itself - either different metrics or source routing or some other vodoo, but I have seen that situation before
2023-08-25 07:34:21 <ikke> telmich: ok. The reason that I noticed is that dns was not working
2023-08-25 07:34:56 <ikke> /etc/resolve.conf was set to .254, but it was not responding to DNS queries 
2023-08-25 07:35:40 <ikke> telmich: thanks for fixing it
2023-08-25 08:36:12 <clandmeter> what happend to gitlab?
2023-08-25 10:08:56 <rnalrd> ikke, I keep getting HTTP500 in gitlab when creating a merge request (after clicking "Compare branches and continue". This happens today, yesterday and the day before.
2023-08-25 10:22:22 <ikke> rnalrd: ok, good to know, haven't noticed it myself
2023-08-25 10:24:42 <ikke> clandmeter: what do you mean, what happened with gitlab?
2023-08-25 10:24:48 <ikke> do you mean those notifications?
2023-08-25 10:24:57 <clandmeter> menu and colors
2023-08-25 10:25:48 <ikke> Well, they change the navigation
2023-08-25 10:26:01 <ikke> https://about.gitlab.com/releases/2023/06/22/gitlab-16-1-released/
2023-08-25 10:27:03 <ikke> and 16.3 will add: https://about.gitlab.com/releases/2023/08/22/gitlab-16-3-released/#new-navigation-has-color-themes-available
2023-08-25 10:37:33 <clandmeter> right, maybe for now we just recolor the logo?
2023-08-25 10:37:39 <clandmeter> looks a bit weird like this
2023-08-25 10:38:11 <ikke> clandmeter: or we wait for 16.3, where it should not be an issue anymore
2023-08-25 10:38:29 <clandmeter> sure, but its like 5 min work? :)
2023-08-25 10:38:43 <ikke> https://gitlab.alpinelinux.org/alpine/infra/infra/-/issues/10807
2023-08-25 10:38:46 <clandmeter> it is fixed white right now?
2023-08-25 10:39:09 <ikke> I think so. I didn't notice because I use the dark theme, so there it looks good
2023-08-25 10:41:07 <clandmeter> ah so its not fixed white, i was refereing to the bg
2023-08-25 10:41:43 <ikke> ok
2023-08-25 10:42:49 <mps> some colors are strange for me also and I use dark theme
2023-08-25 10:43:09 <mps> but that is not important
2023-08-25 10:52:36 <clandmeter> ikke: so the admin menu item is gone now?
2023-08-25 10:52:46 <clandmeter> i think i just saw it but cant find it anymore
2023-08-25 10:53:13 <ikke> It's in the dropdown above 'project overview'
2023-08-25 10:53:42 <clandmeter> oh ok. t hx
2023-08-25 10:54:02 <clandmeter> i added the white logo to the issue
2023-08-25 10:54:12 <clandmeter> i dont see it with white background though :)
2023-08-25 11:34:12 <ikke> iandi: do you already have a gitlab account?
2023-08-25 15:06:24 <iandi> ikke: now yes :) i'm 'usrhere' there
2023-08-25 15:11:16 <ikke> iandi: 👍
2023-08-25 18:23:27 <ptrc> ikke: https://gitlab.alpinelinux.org/alpine/aports/-/jobs/1093838
2023-08-25 18:23:37 <ptrc> ( i imagine a retry will fix that, but letting you know just in case :p )
2023-08-25 18:24:09 <ptrc> ah, it's docker.io that's failing, even better lol
2023-08-25 18:24:16 <ikke> yeah
2023-08-25 18:24:23 <ikke> upstream  issue
2023-08-25 18:24:30 <ptrc> i thought it meant the internal registry, didn't notice the url
2023-08-25 18:25:22 <raspbeguy> Hello
2023-08-25 18:25:45 <raspbeguy> ikke: do we have access to the netbox by now?
2023-08-25 18:26:33 <ikke> I'll arrange it in a sec
2023-08-26 14:55:31 <raspbeguy> Here is an ansible poc that installs zabbix-agent https://gitlab.alpinelinux.org/raspbeguy/ansible-poc
2023-08-26 14:57:26 <raspbeguy> outrageusly simplistic for now. We need to define a config template, so for now I just set a dummy debug message instead of proper config deployment
2023-08-26 17:51:49 <raspbeguy> ikke, clandmeter: ^
2023-08-26 17:52:33 <raspbeguy> We will need to discuss about your zabbix config so I can work on a template 
2023-08-26 18:13:09 <ikke> Thanks for working on it. One thing we also need to discuss is how this would actually get applied (especially in a larger context)
2023-08-26 18:14:29 <ikke> packages: zabbix-agent2, zabbix-agent2-plugin-alpine
2023-08-26 18:14:40 <ikke> /etc/zabbix/zabbix_agent2.conf: https://tpaste.us/b1qL
2023-08-26 18:16:32 <ikke> /etc/zabbix/zabbix_agent2.d/alpine.conf -> ServerActive=zabbix.alpin.pw
2023-08-26 18:18:56 <raspbeguy> Probably by having a AWX server (Ansible Tower community edition)
2023-08-26 18:19:37 <raspbeguy> Or we could also trigger that within gitlab when pushing into master 
2023-08-26 18:20:21 <ikke> This is what clandmeter was working on: https://gitlab.alpinelinux.org/alpine/infra/ansible
2023-08-26 18:23:19 <raspbeguy> I don't have the permission to see this repo
2023-08-26 18:25:23 <ikke> oh, ok, let me fix that
2023-08-26 18:27:58 <ikke> raspbeguy: You should have access now
2023-08-26 18:28:28 <raspbeguy> Yes that's better
2023-08-26 18:33:26 <raspbeguy> So there is already some work on Ansible 
2023-08-26 18:33:34 <clandmeter> That was just a test
2023-08-26 18:33:56 <raspbeguy> So why did you want a poc then?
2023-08-26 18:34:18 <raspbeguy> A poc specifically for zabbix?
2023-08-26 18:36:38 <clandmeter> Something we can extend on
2023-08-26 18:38:26 <raspbeguy> I think you can already extend on what clandmeter did, this is a good start in my opinion 
2023-08-27 14:26:21 <rootwyrm> BTW, ASRock is about to launch a very interesting ARM board. https://www.asrockrack.com/general/productdetail.asp?Model=ALTRAD8UD-1L2T#Specifications
2023-08-27 18:43:15 <clandmeter> ikke: is gitlab down?
2023-08-27 18:44:02 <ikke> checking, don't see anyhting unusual in zabbix
2023-08-27 18:44:10 <ikke> load is high
2023-08-27 18:44:22 <clandmeter> cant load anything
2023-08-27 18:44:43 <ikke> someone making excssive requests again
2023-08-27 18:45:05 <ikke> heh
2023-08-27 18:45:25 <ikke> -> /dev/null
2023-08-27 18:45:26 <clandmeter> didnt gitlab have something to block such calls?
2023-08-27 18:45:36 <ikke> I don't know
2023-08-27 18:46:28 <clandmeter> it loaded now
2023-08-27 18:46:44 <ikke> btw, I use goaccess to quickly identify it
2023-08-27 18:46:55 <ikke> goaccess storage/log/s6/nginx/current
2023-08-27 18:47:10 <ikke> Common log format
2023-08-27 18:47:13 <ikke> section 5
2023-08-27 18:53:27 <clandmeter> https://gitlab.alpinelinux.org/admin/application_settings/network
2023-08-27 18:54:13 <ikke> right
2023-08-27 18:54:17 <clandmeter> https://gitlab.alpinelinux.org/help/user/admin_area/settings/user_and_ip_rate_limits.md
2023-08-27 18:57:14 <clandmeter> ikke: https://alpinelinux.org/ also loads kind of slowish
2023-08-27 18:57:24 <ikke> for me it's instantly
2023-08-27 18:57:34 <clandmeter> yes after some hits it is
2023-08-27 18:57:36 <clandmeter> not sure what it is
2023-08-27 18:57:44 <clandmeter> could also be related to ipv6
2023-08-27 18:58:36 <ikke> That it's faster for me? Because I'm always using ipv6 if possible
2023-08-27 19:21:28 <ayakael> Oh that AsRock board looks quite nice. I've been disappointed with the lack of Arm server parts.
2023-08-27 19:24:23 <ayakael> Although at a grand, gasp. I'm forever stuck with used Xeon boards from 2015...
2023-08-27 20:22:45 <rootwyrm> ayakael: yeah, NGL, my pricing on that board is probably like $475-550. but man, talk about an *easy* bringup and damn fine for production.
2023-08-27 20:23:55 <rootwyrm> Admittedly, I have 'keys to the kingdom' but even then, it's probably gonna be TianoCore and not AMI and should be the Redfish IPMI.
2023-08-27 21:12:51 <rootwyrm> ... oh, teach me to read. AMI 32MB. but the SPI and JTAG are intact, so like... 5 minute bringup. :P
2023-08-27 21:20:35 <ikke> That it's faster for me? Because I'm always using ipv6 if possible
2023-08-27 21:20:38 <ikke> oops
2023-08-28 07:26:29 <mps> 'by mistake' I accepted MR for linux-edge which enabled kexec. should it be reverted or to add patch from linux-lts which disables kexec by default and allow it from kernel cmdline
2023-08-28 13:30:12 <ncopa> I am not able to rebase https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/50663
2023-08-28 13:38:23 <ikke> It's because the branch is master and its protected 
2023-08-28 13:38:35 <ikke> You need to go to the project and unprotect it
2023-08-28 13:39:58 <ncopa> thanks!
2023-08-28 15:07:40 <usrhere> ikke: would you have time, by chance, today to show a bit more around the infra?
2023-08-28 15:09:27 <ikke> usrhere: I'd have more time around thursday
2023-08-28 15:10:01 <usrhere> that would work for me as well
2023-08-28 15:10:13 <usrhere> can we arrange the time already?
2023-08-28 15:14:02 <ikke> I should be available the whole day, but I'm traveling, so I need to make sure I have a working audio setup
2023-08-28 15:16:01 <usrhere> :)
2023-08-28 15:16:52 <usrhere> I could be available any time from 18, on thursday
2023-08-28 15:18:16 <ikke> Ok, should work for me
2023-08-28 15:18:51 <usrhere> nice, thanks a lot
2023-08-28 18:59:15 <raspbeguy> Interested too. Sadly won't be back from work
2023-08-28 19:08:38 <ikke> When will you be back?
2023-08-28 19:21:56 <raspbeguy> When I'm home I then gets busy with my kids... Would 20h be ok?
2023-08-28 19:31:56 <ikke> For me it would 
2023-08-28 20:20:00 <usrhere> ikke: raspbeguy 20 ok for me too
2023-08-28 20:20:14 <ikke> 👍
2023-08-28 20:55:47 <raspbeguy> Great
2023-08-28 21:35:55 <ayakael> By curiosity, what machine does Alpine use for its aarch64 builder?
2023-08-29 08:37:27 <raspbeguy> ikke: I templatized the ansible zabbix poc
2023-08-29 13:47:31 <ncopa> how do I connect to the aarch64 builder machine nowadays?
2023-08-29 13:56:53 <ikke> For edge and 3.18, it's usa-bld-1.alpinelinux.org
2023-08-29 14:18:24 <ncopa> thanks!
2023-08-29 15:23:05 <ikke> lots of spam users created suddenly, all with the same domain
2023-08-29 15:23:25 <ikke> bulk delete :P
2023-08-29 16:54:43 <raspbeguy> ikke: would it be possible to create credentials for me to access zabbix please? Also, creating a test VM for me to test if zabbix agent deployment works would be neat
2023-08-29 17:31:08 <clandmeter> ikke which is the dev box for arm?
2023-08-29 17:31:28 <ikke> che-bld-1
2023-08-29 17:31:47 <clandmeter> builders are on packet?
2023-08-29 17:31:51 <clandmeter> and thx
2023-08-29 17:32:09 <ikke> edge and 3.18
2023-08-29 17:32:16 <ikke> and perhaps 3.17
2023-08-29 17:32:22 <clandmeter> I cannot connect
2023-08-29 17:32:52 <ikke> What are you trying to connect to?
2023-08-29 17:33:14 <clandmeter> my container
2023-08-29 17:33:22 <clandmeter> its off
2023-08-29 17:33:57 <ikke> ah
2023-08-29 17:34:19 <ikke> you disabled autostart for all dev containers :P
2023-08-29 17:34:42 <clandmeter> yup
2023-08-29 17:34:51 <clandmeter> thats how long i didnt play arouind with it
2023-08-29 17:35:07 <ikke> I just started mine the other day as well
2023-08-29 17:35:55 <ikke> raspbeguy: I'll try to set that up tonight
2023-08-29 17:51:30 <raspbeguy> Great
2023-08-30 07:21:41 <clandmeter> ikke: gitlab issues again?
2023-08-30 07:21:48 <clandmeter> loading very slowly
2023-08-30 07:24:41 <ikke> Not slow for me
2023-08-30 07:32:09 <clandmeter> yes is ok now
2023-08-30 07:32:12 <clandmeter> was really slow
2023-08-30 07:44:42 <ikke> I see it in zabbix there were increased error rates 
2023-08-31 09:27:44 <ikke> :/
2023-08-31 09:27:51 <ikke> trying to find out why it's crashing
2023-08-31 15:14:03 <raspbeguy> Will we be meeting in the same room than last time tonight?
2023-08-31 15:22:17 <ikke> Maybe a new room, clandmeter didn't add me to the other room
2023-08-31 21:18:09 <ptrc> oh, is alpine gitlab struggling?
2023-08-31 21:18:21 <ptrc> getting a lot of timeouts when trying to browse merge requests
2023-08-31 21:26:23 <ikke> should be better now
2023-08-31 21:28:34 <ptrc> indeed better :p
2023-08-31 21:28:42 <ptrc> was something actually broken?
2023-08-31 21:28:52 <ptrc> or just a temporary issue
2023-08-31 21:34:26 <ikke> Someone being not so nice 
2023-08-31 22:23:13 <ptrc> also, did we go from 2 jobs per CI runner to just one, or was it always one?