2020-10-01 04:46:31 <ikke> restarted uwsgi for cgit
2020-10-01 05:21:34 <ikke> It was still running, but returned some unspecified error
2020-10-01 08:46:23 <ncopa> im pretty sure uwsgi is doing malloc-after-fork btw. I know because I submitted a PR upstream that does the opendir("/proc/self/fd") thingy. And it was as late as April this year....
2020-10-01 08:47:24 <ikke> ok
2020-10-01 08:47:33 <ikke> But this still uses musl 1.1.24
2020-10-01 13:21:25 <ncopa> does anyone know how to run docker in the lxc dev boxes? I'd like to run something in docker on aarch64 if possible
2020-10-01 13:21:46 <ncopa> this is the error I get:
2020-10-01 13:21:50 <ncopa> docker: Error response from daemon: OCI runtime create failed: container_linux.go:370: starting container process caused: process_linux.go:459: container init caused: process_linux.go:422: setting cgroup config for procHooks process caused: failed to write "c 10:200 rwm" to
2020-10-01 13:21:50 <ncopa> "/sys/fs/cgroup/devices/docker/986d31e2506bd622d8b1becce2b79da7befb3ee862b2a4bf73428921e4bf2120/devices.allow": write /sys/fs/cgroup/devices/docker/986d31e2506bd622d8b1becce2b79da7befb3ee862b2a4bf73428921e4bf2120/devices.allow: operation not permitted: unknown.
2020-10-01 13:22:26 <ikke> ncopa: We have CI VMs that you could use
2020-10-01 13:22:36 <ikke> That's what I typically do to run docker on those arches
2020-10-01 13:23:21 <ikke> ncopa: are you running it inside lxc?
2020-10-01 13:26:12 <ncopa> yes, inside lxc
2020-10-01 13:26:19 <ncopa> and i think i got it working now
2020-10-01 13:26:58 <ncopa> lxc.cgroup{,2}.devices.{deny,allow}
2020-10-01 13:27:12 <ncopa> awesome
2020-10-01 13:28:02 <ikke> cool
2020-10-01 13:28:25 <ncopa> ncopa-edge-aarch64:~$ docker run --rm -it alpine uname -a
2020-10-01 13:28:25 <ncopa> Linux 17dc49843d45 5.4.27-0-lts #1-Alpine SMP Mon, 23 Mar 2020 18:15:20 UTC aarch64 Linux
2020-10-01 13:32:54 <mps> interesting and good news
2020-10-02 07:21:30 <ikke> hmm, somethings fishy
2020-10-02 17:38:07 <Cogitri> Yeah, I think g.a.o isn't happy, some CI builds are stuck on "Getting sources from Git repository"
2020-10-02 17:38:08 <Cogitri> https://gitlab.alpinelinux.org/Cogitri/aports/-/jobs/220748
2020-10-02 18:02:14 <ikke> Cogitri: but that has nothing to do with git.a.o (or at least, should not)
2020-10-02 18:14:12 <Cogitri> Ah right, that should pull from gitlab.a.o
2020-10-02 18:14:20 <Cogitri> Still weird that that's also choking now though
2020-10-02 18:14:32 <Cogitri> Seems like it's just x86_64?
2020-10-02 18:20:51 <ikke> And this is cgit failing, not git over ssh
2020-10-02 18:21:34 <ikke> Ariadne also reported it. 
2020-10-03 11:49:24 <ikke> :-/
2020-10-03 18:16:09 <Cogitri> ikke: Mind pulling asgen's docker-compose repo and restarting it again? :) 
2020-10-04 14:03:10 <Cogitri> ikke: restarted asgen already? :) 
2020-10-05 05:28:47 <ikke> Cogitri: yes (was a bit occupied this weekend)
2020-10-05 05:43:28 <ikke> seems like the uwsgi workers get in some error state, and then they get restarded after 30 minutes of inactivity
2020-10-05 05:43:48 <ikke> uwsgi_proto_uwsgi_parser(): No error information [proto/uwsgi.c line 40]
2020-10-05 05:49:46 <ikke> clandmeter: any idea what might be happening?
2020-10-05 07:10:05 <clandmeter> did something change on that host?
2020-10-05 08:41:32 <ikke> clandmeter: I didn't touch it at least except for restarting uwsgi once when it was stuck
2020-10-05 08:41:36 <ikke> and checking logs
2020-10-06 04:43:00 <ikke> uwsgi didn't get stuck anymore since I last restarted it
2020-10-06 09:01:27 <clandmeter> ok nice
2020-10-06 09:01:35 <clandmeter> still weird
2020-10-06 09:01:38 <ikke> yes
2020-10-06 09:01:42 <ikke> And maybe it will come back
2020-10-06 09:02:02 <clandmeter> i thought it could be related to too many req's
2020-10-06 09:02:39 <ikke> yeah
2020-10-06 12:04:07 <ncopa> https://www.packet.com/ no longer exist. they got bought up by equinix...
2020-10-06 12:05:32 <ikke> right
2020-10-06 12:05:34 <ikke> sadface
2020-10-06 12:05:51 <ikke> Is the console still there?
2020-10-06 12:06:36 <ikke> It forwards to the equinix console
2020-10-06 13:18:34 <clandmeter> looks like the same thing rebranded
2020-10-06 13:30:34 <ikke> indeed
2020-10-06 13:34:53 <clandmeter> heh i just received an email about it
2020-10-06 13:35:01 <clandmeter> an hour ago
2020-10-06 13:38:44 <ikke> Ah yes, me too
2020-10-06 13:38:57 <ikke> Equinix Metal
2020-10-06 13:39:04 <ikke> Sounds like a bandname
2020-10-06 13:39:08 <ikke> Or a music genre
2020-10-06 13:39:13 <clandmeter> :)
2020-10-06 13:50:48 <ikke> clandmeter: uk.alpinelinux.org is still mentioned in the mirror list, but it does no longer resolve
2020-10-06 13:51:25 <ikke> hmm, strange
2020-10-06 13:51:31 <ikke> somewhere else it does resolve
2020-10-06 13:52:04 <ikke> dns resolving is eratic
2020-10-06 13:59:47 <ikke> p.s. Packet is a rad name and a killer url. We’re working on something special. Stay tuned for details. 
2020-10-06 13:59:59 <ikke> Am curious
2020-10-06 14:06:54 <ikke> hmm, build.a.o is down?
2020-10-06 14:07:16 <ikke> Is linode having DNS issues?
2020-10-06 14:07:27 <ikke> yes
2020-10-06 14:07:30 <ikke> https://status.linode.com/
2020-10-06 14:08:24 <ikke> https://status.linode.com/incidents/5k68c46r4sgl
2020-10-06 14:12:26 <ncopa> heh probably related to packet.net -> packet.com change :)
2020-10-06 14:12:52 <ikke> lol
2020-10-06 15:32:25 <ikke> DNS should be working again
2020-10-07 07:28:32 <clandmeter> ikke: what was the problem with uk.a.o ^?
2020-10-07 07:29:08 <clandmeter> ah i already know
2020-10-07 07:34:05 <clandmeter> nope i dont :)
2020-10-07 07:34:30 <clandmeter> ikke: any clue?
2020-10-07 09:01:11 <ikke> clandmeter: linode
2020-10-07 09:01:16 <ikke> but
2020-10-07 09:01:24 <ikke> we redirect uk.a.o to nld3
2020-10-07 13:33:34 <clandmeter> ikke: linode?
2020-10-07 13:33:46 <ikke> linode had a DNS issue yesterday
2020-10-07 13:33:59 <ikke> Or are you referring to Zabbix complaining about uk?
2020-10-07 13:34:08 <clandmeter> yes
2020-10-07 13:34:43 <ikke> "Preprocessing failed for: /usr/lib/zabbix/externalscripts/check_http: line 38: curl: not found.env: can't execute 'python3'..."
2020-10-07 13:35:15 <ikke> didn't look much further into it yet
2020-10-07 13:43:33 <clandmeter> thats only for uk.a.o?
2020-10-07 13:44:34 <ikke> From what I can see, yes
2020-10-08 13:21:31 <ikke> clandmeter: would be nice if we could automatically get tickets in gitlab for these mirror requests
2020-10-08 13:27:53 <clandmeter> those mirrors are already added i think
2020-10-08 13:27:57 <clandmeter> we need to inform them
2020-10-08 13:28:08 <clandmeter> and yes, added to tickets would be nice
2020-10-08 17:37:35 <ikke> clandmeter: working on a 13.2 image for gitlab
2020-10-08 17:37:44 <ikke> what should I take into account?
2020-10-08 17:39:21 <ikke> (I've upgraded gitlab.a.o to 13.1.11)
2020-10-09 08:04:56 <ikke> clandmeter: reading here that the new bigger ARM cores will no longer support 32-bits (from 2022)
2020-10-09 08:21:40 <clandmeter> ikke: haha
2020-10-09 08:21:48 <clandmeter> i was just going to mention that here
2020-10-09 08:22:06 <clandmeter> well i wanted to do that yesterday but got too busy...
2020-10-09 08:22:23 <clandmeter> so we need to keep a few older machines around for 32bit
2020-10-09 08:22:33 <ikke> heh
2020-10-09 08:22:42 <ikke> Just read it on a dutch tech new site (tweakers)
2020-10-09 08:22:52 <clandmeter> yes same here
2020-10-09 08:23:09 <clandmeter> tab is still open :)
2020-10-09 08:23:13 <ikke> heh
2020-10-09 13:17:06 <HRio> is the ppc64le builder in CI always slow on artifact upload (compared to the other builders)?
2020-10-09 13:17:19 <ikke> Yes
2020-10-09 13:17:48 <HRio> ah OK
2020-10-09 13:17:48 <ikke> Some kind of network issue that is slowing down these uploads. Some times it's worse than other times
2020-10-09 13:20:34 <HRio> seems its about same performance when it pulls down stuff as the others
2020-10-09 13:21:01 <HRio> but the push is very very slow minutes for < 50MB
2020-10-09 13:21:22 <ikke> yes, upload speed is affected onyl
2020-10-09 13:22:16 <HRio> k
2020-10-11 06:14:47 <ikke> 
ACTION should look at that

2020-10-12 08:09:37 <ncopa> morning! I have an issue with bluez update, which has a failing test on x86_64 (only). It looks like the test needs some specific kernel config to pass: https://github.com/bluez/bluez/issues/40
2020-10-12 08:10:22 <ncopa> I think we have two options: 1) disable the test (we should be able to build it regardless of kernel config) 2) update the kernel on the host
2020-10-12 08:10:34 <ncopa> does infra team has any opinion on this?
2020-10-12 08:25:35 <ikke> It's already on 3.12. Do you plan backporting these config changes?
2020-10-12 09:59:02 <ncopa> $ uname -a
2020-10-12 09:59:03 <ncopa> Linux ncopa-edge-x86_64 4.19.80-0-vanilla #1-Alpine SMP Fri Oct 18 11:27:53 UTC 2019 x86_64 GNU/Linux
2020-10-12 09:59:07 <ncopa> maybe it just needs a reboot
2020-10-12 09:59:51 <ikke> /lib/modules and uname -r agreen with eachother
2020-10-12 10:05:38 <ncopa> maybe we need to migrate from -vanilla to -lts kernel?
2020-10-12 10:09:38 <ikke> right
2020-10-12 10:09:55 <ikke> /boot contains -lts
2020-10-12 10:10:00 <ncopa> i wonder if i should backport the kernel updates to 3.12 first
2020-10-12 10:10:36 <ikke> update-extlinux.conf points to lts
2020-10-12 10:10:56 <ncopa> $ uptime
2020-10-12 10:10:56 <ncopa>  12:10:47 up 313 days, 23:28,  load average: 3.32, 3.96, 3.51
2020-10-12 10:11:00 <ncopa> may need a reboot
2020-10-12 10:11:24 <ikke> which server is that?
2020-10-12 10:11:35 <ikke> 10:11:08 up 97 days, 21:12,  load average: 6.97, 22.43, 19.64
2020-10-12 10:11:40 <ncopa> x86_64 machine
2020-10-12 10:11:54 <ikke> I'm looking at nld9-dev1
2020-10-12 10:11:55 <ncopa> nld5-dev1 iirc
2020-10-12 10:11:57 <ikke> aha
2020-10-12 10:12:02 <ikke> the dev machine
2020-10-12 10:12:13 <ikke> not the builders
2020-10-12 10:12:34 <ncopa> oh, right, they are on different machines. sorry
2020-10-12 10:13:03 <ikke> That one is still alpine 3.10
2020-10-12 10:13:11 <ikke> And I think we should make effort to upgrade our infra to 3.12
2020-10-12 12:21:01 <ncopa> im updating the kernel for 3.12 now
2020-10-12 12:21:54 <ikke> ok
2020-10-15 23:09:13 <Ariadne> build.alpinelinux.org DNS queries return SERVFAIL
2020-10-15 23:14:16 <Ariadne> seems to have recovered
2020-10-16 04:46:33 <ikke> hmm, more linode dns issues?
2020-10-16 04:47:06 <ikke> yup: https://status.linode.com/incidents/wssx0cf2dbhb
2020-10-17 10:23:43 <Cogitri> The armv7 runner seems very unhappy
2020-10-17 10:24:15 <ikke> whats up
2020-10-17 10:24:18 <Cogitri> https://gitlab.alpinelinux.org/Cogitri/aports/-/jobs/229755
2020-10-17 10:24:39 <ikke> Seems to be hanging on `sync`
2020-10-17 10:25:34 <ikke> Oh you are talking about the CI 
2020-10-17 10:25:41 <ikke> The builder seems to be unhappy as well
2020-10-17 10:27:01 <Cogitri> Oh :/ 
2020-10-17 10:30:14 <ikke> note they are on the same hardware, so could be related
2020-10-17 10:31:14 <ikke> going to reboot the host
2020-10-17 10:34:37 <mps> usa4?
2020-10-17 10:35:11 <ikke> yes
2020-10-17 10:35:36 <mps> huh
2020-10-17 10:35:37 <ikke> But even the reboot command is not working :/
2020-10-17 10:36:00 <ikke> qemu is in an uninterruptable sleep
2020-10-17 10:36:51 <mps> load is about 14
2020-10-17 10:37:23 <mps> for me it works fine
2020-10-17 10:38:04 <ikke> [10056580.556282] blk_update_request: I/O error, dev sdb, sector 64275984 op 0x1:(WRITE) flags 0x4a00 phys_seg 5 prio class 0
2020-10-17 10:38:09 <ikke> [10056580.556292] sd 3:0:0:0: rejecting I/O to offline device
2020-10-17 10:38:25 <ikke> rebooting it now
2020-10-17 10:38:36 <ikke> mps: okay with you?
2020-10-17 10:38:40 <mps> uh, can you wait a minute
2020-10-17 10:38:43 <ikke> yes
2020-10-17 10:39:33 <mps> ok, go now
2020-10-17 10:40:00 <mps> I hope it will sync disk
2020-10-17 10:41:02 <ikke> it should
2020-10-17 10:41:04 <mps> iscsi again?
2020-10-17 10:41:06 <ikke> I think so
2020-10-17 10:44:08 <ikke> Loading initial ramdisk ..
2020-10-17 10:44:17 <mps> algitbot: /o\
2020-10-17 10:45:05 <ikke> ok, it's up now
2020-10-17 10:49:44 <ikke> Cogitri: the job runs now, but fails because the branch is gone
2020-10-17 10:53:49 <mps> From 172.16.0.12 (172.16.0.12) icmp_seq=1 Destination Host Unreachable
2020-10-17 10:55:21 <ikke> 64 bytes from 172.16.0.12: icmp_seq=1 ttl=63 time=86.0 ms 
2020-10-17 10:57:36 <mps> monte carlo don't like me today ;)
2020-10-17 10:58:53 <mps> https://tpaste.us/qgow
2020-10-17 11:00:04 <ikke> https://tpaste.us/N87k
2020-10-17 11:05:52 <ikke> mps: Do you have issues connecting to your container?
2020-10-17 11:06:11 <mps> yes
2020-10-17 11:06:32 <ikke> What IP?
2020-10-17 11:06:40 <mps> with vpn and with ssh port
2020-10-17 11:07:07 <mps> sh: connect to host usa4.alpinelinux.org port xxxxx: Host is unreachable
2020-10-17 11:07:15 <mps> ssh*
2020-10-17 11:08:20 <Cogitri> ikke: Ok thanks for looking into it 👍
2020-10-17 11:08:21 <ikke> I don't see any ports forwarded to your container
2020-10-17 11:08:22 <mps> 'ping usa4.alpinelinux.org' works
2020-10-17 11:08:49 <mps> it doesn't work with vpn also
2020-10-17 11:09:05 <ikke> mps: which container, you have 2
2020-10-17 11:09:11 <mps> and till reboot ssh portfwd worked
2020-10-17 11:09:20 <mps> both
2020-10-17 11:09:33 <ikke> what IP?
2020-10-17 11:09:48 <mps> 147.75.79.110
2020-10-17 11:09:54 <ikke> internal?
2020-10-17 11:10:06 <mps> wait moment
2020-10-17 11:10:27 <ikke> They might have gotten a different IP address
2020-10-17 11:10:39 <mps> 172.16.12.10
2020-10-17 11:10:43 <ikke> ah, yes
2020-10-17 11:11:04 <ikke> dhcp gave them different addresses
2020-10-17 11:11:15 <ikke> which container was that?
2020-10-17 11:11:30 <ikke> I see
2020-10-17 11:11:47 <mps> huh, mps-aarch64- like this
2020-10-17 11:12:09 <ikke> I think that's the ifupdown-ng lack of supporting hostname $(hostname)
2020-10-17 11:12:23 <mps> hmmm
2020-10-17 11:12:47 <mps> that is why I don't use ifupdown-ng
2020-10-17 11:13:04 <mps> and don't think it should be default
2020-10-17 11:13:12 <ikke> meh, just small issues that will be fixed
2020-10-17 11:14:14 <mps> heh, where is time when pkgs are kept in testing at least one release cycle
2020-10-17 11:27:28 <ikke> Ariadne: this should still work with ifupdown-ng, right? hostname mps-edge-aarch64
2020-10-17 11:27:53 <Ariadne> yes
2020-10-17 11:28:05 <ikke> for some reason I keep getting a dynamic address
2020-10-17 11:28:27 <Ariadne> is it latest ifupdown-ng?
2020-10-17 11:28:55 <ikke> let me check
2020-10-17 11:29:23 <Ariadne> it'd be nice if we could look into matrix again btw
2020-10-17 11:29:30 <ikke> No, apparently not
2020-10-17 11:29:37 <Ariadne> they finally have a homeserver that is better than synapse
2020-10-17 11:29:40 <ikke> Installed:                                Available:
2020-10-17 11:29:43 <ikke> ifupdown-ng-0.8.5-r0                    < 0.10.0-r1
2020-10-17 11:29:51 <Ariadne> yeah 0.9 and newer support hostname
2020-10-17 11:29:52 <Ariadne> :)
2020-10-17 11:29:55 <ikke> ok
2020-10-17 11:30:02 <ikke> mps: mind if I upgrade ifupdown-ng in your containers?
2020-10-17 11:30:19 <Ariadne> it was an unintentional omission
2020-10-17 11:30:58 <Ariadne> 0.10 features lots of stuff
2020-10-17 11:31:03 <ikke> And the latest version has that config option as well, right?
2020-10-17 11:31:21 <Ariadne> which config option
2020-10-17 11:31:26 <Ariadne> hostname?
2020-10-17 11:31:27 <Ariadne> yes
2020-10-17 11:31:41 <ikke> yes, hostname from uname
2020-10-17 11:31:43 <mps> ikke: ok
2020-10-17 11:32:40 <ikke> ok, fixed
2020-10-17 11:32:42 <ikke> should work now
2020-10-17 11:33:07 <mps> it works now
2020-10-17 11:34:17 <mps> Ariadne: why CoC in ifupdown-ng
2020-10-17 11:34:18 <Ariadne> nice
2020-10-17 11:34:43 <mps> ofc, ypu are free to do whatever you want
2020-10-17 11:34:53 <Ariadne> $dayjob
2020-10-17 11:34:58 <mps> you*
2020-10-17 11:35:09 <mps> ah, understand
2020-10-17 11:35:27 <Ariadne> $dayjob actually thinks it is quite silly, but does not want to deal with it
2020-10-17 11:35:47 <mps> meh, what to say
2020-10-17 11:35:54 <Ariadne> so we install a CODE_OF_CONDUCT.md file containing a CoC we would reasonably enforce anyway
2020-10-17 11:36:24 <mps> my aversion to ifupdown-ng now vanished :)
2020-10-17 11:36:52 <Ariadne> i mean, if some asshole shows up doing fucked up shit
2020-10-17 11:36:55 <Ariadne> i am going to ban them
2020-10-17 11:37:01 <Ariadne> regardless of the presence of that file
2020-10-17 11:37:09 <mps> sure
2020-10-17 11:37:13 <Ariadne> but you can't just say that and call it a day anymore
2020-10-17 11:38:14 <mps> if I ever made CoC I would ban 'fu.k*' and 'sh.t' words :D
2020-10-17 11:38:54 <Ariadne> anyway i would seriously like to propose we take another look at matrix
2020-10-17 11:39:00 <Ariadne> not because i love matrix, but because i hate freenode
2020-10-17 11:39:03 <Ariadne> )
2020-10-17 11:39:06 <Ariadne> :)*
2020-10-17 11:41:58 <Ariadne> mps: but yeah, since $dayjob pays me to work on ifupdown-ng (and alpine in general), they don't want drama with concern trolls
2020-10-17 11:42:10 <Ariadne> so CoC gets slapped on all projects as policy
2020-10-17 11:42:47 <mps> understand (to repeat)
2020-10-17 11:44:16 <mps> maybe I'm naive, but I think people can communicate and work without written and enforced rules
2020-10-17 11:44:33 <mps> if we need rules better to stop then
2020-10-17 11:46:23 <Ariadne> i agree
2020-10-17 11:46:42 <Ariadne> healthy projects -- generally don't need something like a CoC
2020-10-17 11:46:49 <Ariadne> unless they grow to be very large
2020-10-17 11:46:56 <mps> heh
2020-10-17 11:46:59 <mps> yes
2020-10-17 11:47:08 <Ariadne> in which case the CoC is useful because it sets the expectations that moderators would use to moderate
2020-10-19 12:49:31 <ncopa> ikke: re alpine-mksite, do you know how to get the "production" branch published?
2020-10-19 12:50:01 <ncopa> i just published an orange icon that i was thinking we could use for icon for our gitlab
2020-10-19 12:57:11 <ikke> usually when the production branch was updated, it would do that
2020-10-19 13:19:29 <ikke> ncopa: seems like you already changed it for gitlab.a.o?
2020-10-19 13:22:25 <ncopa> yeah. it looks like you simply upload it to gitlab
2020-10-19 13:22:28 <ncopa> what do you think :)
2020-10-19 13:23:34 <ikke> nice to be able to distinguish gitlab tabs from other alpine tabs :)
2020-10-19 17:27:52 <ikke> clandmeter: https://gitlab.alpinelinux.org/alpine/infra/docker/gitlab/-/commits/13.2-stable
2020-10-19 17:31:51 <ikke> https://cloud.drone.io/alpinelinux/alpine-docker-gitlab/138
2020-10-19 18:43:20 <clandmeter> good evening
2020-10-19 18:43:48 <clandmeter> sorry, been a hectic day(s)
2020-10-19 18:46:31 <ikke> clandmeter: I kind of assumed that
2020-10-19 18:46:54 <ikke> clandmeter: I'm restoring a gitlab backup now
2020-10-19 18:48:07 <clandmeter> well im really sorry most of infra is now on your shoulders, i would prefer to see it differently. but at this moment its what it is.
2020-10-19 18:48:15 <clandmeter> but im here if you have questions
2020-10-19 18:48:51 <clandmeter> and i dont see it changing in a near future.
2020-10-19 18:49:07 <mps> clandmeter: \o/
2020-10-19 18:49:44 <clandmeter> would be nice to have somebody support infra
2020-10-19 18:49:55 <clandmeter> 
ACTION hugs mps 

2020-10-19 18:50:16 <clandmeter> its a digital one, so should be safe :)
2020-10-19 18:50:24 <clandmeter> (i know you dont mind anyway)
2020-10-19 18:50:35 <mps> :)
2020-10-19 18:50:52 <clandmeter> mps: how is life in your area?
2020-10-19 18:51:05 <mps> yes, I don't care about this histeria
2020-10-19 18:51:22 <mps> well, mostly fine
2020-10-19 18:51:47 <clandmeter> everyting closed?
2020-10-19 18:51:55 <mps> not yet
2020-10-19 18:52:10 <mps> and I hope it will not be
2020-10-19 18:52:27 <ikke> mps: do you call more than 1 milion people dying, histeria? :)
2020-10-19 18:52:33 <clandmeter> i just went to the city center 
2020-10-19 18:52:48 <clandmeter> get new glasses
2020-10-19 18:53:14 <clandmeter> looked a bit sad with all things closed
2020-10-19 18:53:16 <mps> last time government tried to make 'curfew' were big demonstrations and they 'ceased'
2020-10-19 18:53:30 <ikke> people are selfish
2020-10-19 18:53:45 <clandmeter> people are a lot of things :)
2020-10-19 18:54:00 <mps> ikke: last year 1.7 milion peoples died of flu
2020-10-19 18:54:57 <mps> every day children dies in world in not small numbers because they don't have enough food.
2020-10-19 18:55:05 <mps> but we are OT
2020-10-19 18:55:22 <ikke> :)
2020-10-19 18:56:32 <mps> clandmeter: DHL 'boy' here about month ago but with airplane, so I couldn't bother him with bottle :( 
2020-10-19 18:56:43 <mps> was here*
2020-10-19 18:56:58 <clandmeter> mps: offtopic channel :p
2020-10-19 18:57:06 <mps> sure
2020-10-19 18:58:05 <ikke> ontopic: anythign I should verify when updating the gitlab image?
2020-10-19 18:58:24 <clandmeter> didnt we have a test suite?
2020-10-19 18:58:42 <ikke> an acceptance test suite, after it has been deployed
2020-10-19 18:58:46 <ikke> which is quite basic
2020-10-19 19:00:27 <clandmeter> right
2020-10-19 19:00:35 <clandmeter> well we didnt really working on anything else iirc
2020-10-19 19:01:14 <ikke> I'll deploy the current image and see how it goes
2020-10-19 19:01:26 <clandmeter> did you check the changelog
2020-10-19 19:01:29 <clandmeter> and diffs?
2020-10-19 19:01:36 <ikke> Not really
2020-10-19 19:01:44 <clandmeter> most of the items i bump into are config changes
2020-10-19 19:01:48 <ikke> Yes
2020-10-19 19:01:50 <ikke> I fixed one
2020-10-19 19:01:54 <ikke> at least, where a patch conflicted
2020-10-19 19:02:01 <clandmeter> logrotate
2020-10-19 19:02:05 <ikke> yup
2020-10-19 19:02:21 <ikke> I manually updated the patch :P
2020-10-19 19:03:06 <clandmeter> i still didnt hear anything back from linode
2020-10-19 19:03:10 <ikke> Had to reboot usa4-dev3 again btw
2020-10-19 19:03:37 <clandmeter> i dont have much time, so im not that pushy 
2020-10-19 19:03:46 <ikke> We'll see
2020-10-19 19:04:07 <clandmeter> LBlaboon: do you have any news?
2020-10-19 19:04:22 <clandmeter> just to keep the conversation going :)
2020-10-19 21:47:36 <LBlaboon> clandmeter: apologies for the silence, i actually just got back from 2 weeks vacation. i'm actually working with management on our end and i think they're drafting up a formal policy that would bump your sponsorship to $1000/month in exchange for two "promotional activities" (interviews and such) per year, with this year's activities already being completed
2020-10-19 21:56:00 <LBlaboon> i'll try to see how far along they are with that
2020-10-20 18:35:44 <ikke> clandmeter: got gitlab running on gitlab-test, but the API is again not available
2020-10-20 18:35:49 <ikke> Will look into
2020-10-20 18:35:52 <ikke> it
2020-10-21 06:09:56 <ncopa> morning! I'd like to reboot nld5-dev1
2020-10-21 06:10:46 <ncopa> Linux nld5-dev1 4.19.80-0-vanilla is running, but 
2020-10-21 06:10:47 <ncopa> nld5-dev1 [~]# ls /lib/modules/
2020-10-21 06:10:47 <ncopa> 4.19.118-0-vanilla  4.9.65              backup
2020-10-21 06:10:58 <ncopa> 4.19.118 is installed
2020-10-21 06:11:41 <mps> ah
2020-10-21 06:11:43 <ncopa> im trying to load some kernel modules for https://github.com/bluez/bluez/issues/40#issuecomment-705074798
2020-10-21 06:11:57 <ikke> ncopa: Do we want to upgrade nld5 to 3.12 as well?
2020-10-21 06:12:14 <mps> I wanted to ask ikke few days ago to upgrade kernel to -lts
2020-10-21 06:12:25 <ncopa> we could....
2020-10-21 06:12:30 <ikke> ncopa: it's on my todo list
2020-10-21 06:13:17 <mps> when it will be rebooted? I want to stop postgresql there before that
2020-10-21 06:13:41 <ikke> mps: shouldn't postgres be able to deal with shutdowns?
2020-10-21 06:14:02 <ikke> or are you running something?\
2020-10-21 06:14:07 <mps> it should on native, not sure if it runs in lxc
2020-10-21 06:14:23 <ikke> lxc is native..
2020-10-21 06:14:44 <mps> then ok, do whenever you want
2020-10-21 09:48:04 <ikke> ncopa: ^
2020-10-21 09:54:29 <ncopa> looks like our mirrors are running out of disk space
2020-10-21 09:55:02 <ikke> yes
2020-10-21 09:56:32 <ncopa> im deleteing the release candidates on the dl-master
2020-10-21 09:58:50 <ncopa> that should have free'd up ~62G
2020-10-21 09:59:12 <ncopa> we may need do something for the 3.13 release
2020-10-21 10:05:55 <ikke> that at least helped
2020-10-21 10:25:25 <ncopa> hum....
2020-10-21 10:25:48 <ncopa> i think i may have a problem with my gpg key for signing the 3.12.1 release
2020-10-21 10:26:25 <ncopa> gpg: skipped "07D9495A": secret key not available
2020-10-21 10:26:25 <ncopa> gpg: signing failed: secret key not available
2020-10-21 10:28:37 <ikke> :o
2020-10-21 10:28:52 <ncopa> where did that key go?
2020-10-21 10:30:47 <ikke> 
ACTION hopes that key is backed up somewhere

2020-10-21 10:32:12 <ncopa> i do have the key backed up
2020-10-21 10:32:14 <ncopa> i think
2020-10-21 10:32:20 <mps> paper trail
2020-10-21 10:32:37 <ncopa> but i wonder where it went?
2020-10-21 10:32:49 <ncopa> did the key use some algo that no longer supported?
2020-10-21 10:33:03 <mps> no, afaik
2020-10-21 10:33:18 <mps> I have old keys and they still works
2020-10-21 10:33:27 <ncopa> i have a copy of the ~/.gnupg when i restore from backup the same error happens
2020-10-21 10:35:12 <mps> I have .gnupg/private-keys-v1.d dir with more keys
2020-10-21 10:35:47 <ncopa> how do i know which is which
2020-10-21 10:37:15 <mps> it should be first 8 characters in filename
2020-10-21 10:38:08 <mps> huh, pinentry doesn't work on edge for me
2020-10-21 10:41:51 <mps> uh, and gpg doesn't list secret keys properly. have to boot 3.12 and check
2020-10-21 10:45:06 <ncopa> gnupg from alpine 3.12 does list the secret keys
2020-10-21 10:45:43 <ikke> so a recent gpg update
2020-10-21 10:46:35 <ikke> 2.2.21 and 2.2.23
2020-10-21 10:46:54 <ncopa> huh
2020-10-21 10:47:04 <ikke> hmm
2020-10-21 10:47:09 <ikke> 3.12 has 2.2.23
2020-10-21 10:47:13 <ncopa> $ gpg --version
2020-10-21 10:47:13 <ncopa> gpg (GnuPG) 1.4.23
2020-10-21 10:47:20 <ncopa> i think that may be the problem :)
2020-10-21 10:47:21 <ikke> ah, gnupg1.
2020-10-21 10:47:28 <ncopa> i had gnupg1 :)
2020-10-21 10:47:43 <ikke> explains
2020-10-21 10:49:46 <ncopa> (1/1) Reinstalling gnupg1 (1.4.23-r0)
2020-10-21 10:49:56 <ncopa> that is what apk fix gnupg does
2020-10-21 10:52:45 <ikke> seems like both gnupg1 as gnupg2 provide gnupg?>
2020-10-21 10:53:04 <ikke> yes
2020-10-21 10:53:41 <ikke> gnupg1 has provides gnupg=1.4.23-r0
2020-10-21 10:54:05 <ncopa> i cannot install gnupg2 nor delete gnupg1
2020-10-21 10:54:13 <ikke> there is no gnupg2
2020-10-21 10:54:18 <ikke> there is gnupg1 and gnupg
2020-10-21 10:54:25 <ncopa> gnupg>2
2020-10-21 10:54:29 <ikke> right
2020-10-21 10:54:33 <ncopa> ERROR: unable to select packages:
2020-10-21 10:54:33 <ncopa>     breaks: world[gnupg>2]
2020-10-21 10:54:33 <ncopa>   gnupg1-1.4.23-r0:
2020-10-21 10:54:33 <ncopa>     satisfies: lxc-download-4.0.4-r0[gnupg1] gpgme-1.14.0-r0[gnupg]
2020-10-21 10:54:33 <ncopa>                gpgmepp-1.14.0-r0[gnupg] dns-root-hints-2019073000-r2[gnupg]
2020-10-21 10:54:37 <ikke> so does something depend on gnupg <2?
2020-10-21 10:54:43 <ncopa> i think it may be due to the missing libreoffice packages?
2020-10-21 10:54:50 <ncopa> World updated, but the following packages are not removed due to:
2020-10-21 10:54:50 <ncopa>   gnupg1: lxc-download gpgme gpgmepp libreoffice-common libreoffice-calc
2020-10-21 10:54:50 <ncopa>           libreofficekit libreoffice libreoffice-base
2020-10-21 10:54:50 <ncopa>           libreoffice-connector-postgres libreoffice-writer libreoffice-gnome
2020-10-21 10:54:51 <ncopa>           libreoffice-draw libreoffice-math libreoffice-impress
2020-10-21 10:54:51 <ncopa>           claws-mail-plugins-pgp claws-mail-plugins-smime volume_key
2020-10-21 10:54:51 <ncopa>           libblockdev udisks2 dns-root-hints unbound
2020-10-21 10:55:25 <ncopa> $ apk version -l '?' | tpaste
2020-10-21 10:55:25 <ncopa> https://tpaste.us/Yovg
2020-10-21 10:55:25 <ncopa> WARNING: Ignoring /home/ncopa/packages/non-free: No such file or directory
2020-10-21 10:56:01 <mps> yes, dependency hell, like debian
2020-10-21 10:56:41 <ikke> Would be nice if there was some nicer way to deal with these issues
2020-10-21 10:56:57 <ikke> on major upgrades, we tend to get a lot of users with these kinds of problems
2020-10-21 10:57:29 <ncopa> how about drop gnupg1?
2020-10-21 10:58:16 <mps> I fear a lot of people will start to bashing
2020-10-21 11:03:39 <ncopa> what if we drop the provides?
2020-10-21 11:04:04 <ikke> ncopa: Sounds better
2020-10-21 11:07:20 <mps> maybe move gnupg1 to testing
2020-10-21 11:07:30 <ikke> Does not make a lot of sense to me
2020-10-21 11:07:33 <ikke> to move packages back to testing
2020-10-21 11:08:03 <mps> for those who need it for some backward compatibility?
2020-10-21 11:08:17 <ikke> still doesn't make a lot of sense
2020-10-21 11:08:40 <mps> personally I'm for removing it
2020-10-21 11:08:44 <ikke> We can move it to community, or unmaintaned, but I dont think it makes a lot of sense to move things back to testing
2020-10-21 11:09:49 <mps> aha, create 'obsolete' subdir in aports repo
2020-10-21 11:12:32 <mps> oh, good pinentry-ncurses works on edge
2020-10-25 17:08:29 <ikke> Fun deadlocks again in gitlab/gitaly
2020-10-27 07:40:06 <ikke> jkjj\o/
2020-10-27 09:10:31 <clandmeter> ikke: how are things?
2020-10-27 09:11:33 <ikke> Personally I'm doing fine :)
2020-10-27 09:25:21 <ikke> clandmeter: With you?
2020-10-27 10:49:46 <clandmeter> busy, but i guess thats obvious :)
2020-10-27 10:49:52 <ikke> :)
2020-10-27 10:49:57 <clandmeter> what about infra?
2020-10-27 10:49:59 <ikke> I still need to fix the gitlab image
2020-10-27 10:50:01 <clandmeter> all ok?
2020-10-27 10:52:57 <clandmeter> the deadlock issue is related to musl?
2020-10-27 10:53:01 <clandmeter> musl/ruby
2020-10-27 10:53:37 <ikke> I would say so
2020-10-27 10:53:46 <ikke> same symptoms
2020-10-27 10:54:44 <clandmeter> did ncopa report it upsteam?
2020-10-27 10:55:05 <ikke> He did report this: https://bugs.ruby-lang.org/issues/17189
2020-10-27 10:55:54 <clandmeter> looks active...
2020-10-27 10:56:20 <ikke> yup..
2020-10-27 11:35:50 <ncopa> dalias posted an experimental patch to musl recently, that would make malloc after fork in thread work
2020-10-27 11:36:12 <ncopa> i think we should test his patch
2020-10-27 11:36:41 <ikke> In our gitlab instance?
2020-10-27 11:39:39 <mps> better test it somewhere less critical than on infra service, if someone have time
2020-10-27 11:39:50 <ikke> heh :)
2020-10-27 11:41:04 <mps> maybe I can find time this night
2020-10-27 11:41:18 <mps> what should be good test
2020-10-27 13:19:45 <clandmeter> we could run it first in our test instance
2020-10-27 13:19:51 <ikke> Tyup
2020-10-27 13:19:53 <clandmeter> see if general operation works
2020-10-27 13:19:54 <ikke> but it has no load
2020-10-27 13:20:11 <clandmeter> if the patch works, the general functions will work
2020-10-27 13:20:24 <clandmeter> its worth a try in production
2020-10-27 13:20:30 <ikke> nod
2020-10-27 13:20:34 <clandmeter> and docker can revert fast :)
2020-10-27 13:20:57 <clandmeter> but i guess we firest need to find the issue you are having now
2020-10-27 13:21:00 <ikke> yes, could create a tagged image for that
2020-10-27 13:21:13 <ikke> Yes, something with protobufs again I guess
2020-10-27 13:21:18 <ikke> internal API not available
2020-10-27 13:21:36 <ikke> But I did not investigate deeply (ie, compare versions, etc)
2020-10-27 13:22:45 <clandmeter> do you have a strace?
2020-10-27 13:22:53 <clandmeter> or err msg
2020-10-27 13:26:34 <ikke> https://tpaste.us/lEZo 
2020-10-27 13:27:47 <clandmeter> right
2020-10-27 13:27:57 <clandmeter> i think there is a more details error in gitlab logs
2020-10-27 13:28:07 <clandmeter> production.log or similar
2020-10-27 13:28:10 <ikke> Didn't see sidekiq had issues too
2020-10-28 20:04:12 <ikke> https://news.ycombinator.com/item?id=24919569 has sime threads about spam issues with gitlab (and potential solutions)
2020-10-28 20:04:17 <ikke> clandmeter: ^
2020-10-28 20:06:08 <ikke> https://gitlab.com/gitlab-org/gitlab/-/issues/273480
2020-10-29 14:57:54 <ncopa> how can i add a gitlab runner for abuild project?
2020-10-29 14:58:37 <ncopa> https://gitlab.alpinelinux.org/alpine/abuild/-/merge_requests/64
2020-10-29 15:00:19 <ikke> https://gitlab.alpinelinux.org/alpine/abuild/-/settings/ci_cd
2020-10-29 15:00:31 <ikke> Under runners yuo can see what runners are available
2020-10-29 15:00:40 <ikke> The trick is to use the right tags 
2020-10-29 15:01:09 <ikke> Oh, you need to 'enable shared runners'
2020-10-29 15:03:18 <ncopa> i'll have a look at that other day
2020-10-29 15:04:07 <ikke> In the jobs here (check, build): https://gitlab.alpinelinux.org/alpine/abuild/-/merge_requests/64/diffs
2020-10-29 15:04:20 <ikke> you need to add tags: ['docker-alpine']
2020-10-29 15:11:24 <ncopa> humpf... creating releases in gitlab is not as simple as create a tag
2020-10-29 15:14:46 <ikke> ncopa: isn't that just a matter of adding release notes additionally to the tag message?
2020-10-29 15:17:22 <ncopa> does not look like i can download the tarballs with curl. i get http 500
2020-10-29 15:22:26 <ikke> which tarballs?
2020-10-29 15:33:03 <ncopa> https://gitlab.alpinelinux.org/alpine/abuild/-/archive/3.7.0_rc1/abuild-3.7.0_rc1.tar.bz2
2020-10-29 15:33:24 <ncopa> abuild was not happy at least
2020-10-29 15:34:46 <ikke> hmm, bz2
2020-10-30 12:10:15 <clandmeter> ncopa: where is this musl patch?
2020-10-30 12:11:10 <ikke> clandmeter: https://www.openwall.com/lists/musl/2020/10/28/6/1
2020-10-30 12:11:14 <ikke> https://www.openwall.com/lists/musl/2020/10/28/6
2020-10-30 12:11:57 <mps> mt-fork?
2020-10-30 12:12:55 <clandmeter> ikke: current alpine branch for gitlab is 3.12?
2020-10-30 12:13:03 <ikke> yes
2020-10-30 12:13:22 <mps> here is what I made for alpine https://tpaste.us/qgKe
2020-10-30 12:13:22 <ikke> so.. question is why we even have this issue
2020-10-30 12:13:51 <mps> oh, gitlab, sorry
2020-10-30 12:14:51 <clandmeter> ikke: ?
2020-10-30 12:15:07 <clandmeter> this patch is already in 3.12?
2020-10-30 12:15:10 <ikke> the issue mostly came up with the latest musl version, right?
2020-10-30 12:15:35 <clandmeter> you were the first to notice it
2020-10-30 12:15:41 <clandmeter> i think when you switched to 3.12
2020-10-30 12:19:51 <clandmeter> mps: why are you removing those other patches in your diff?
2020-10-30 12:20:28 <mps> they are already in master
2020-10-30 12:20:49 <mps> Ariadne took them from master and added
2020-10-30 12:20:56 <clandmeter> pkgver didnt change?
2020-10-30 12:21:02 <ikke> no
2020-10-30 12:21:05 <ikke> only pkgrel
2020-10-30 12:21:09 <mps> no but -r4
2020-10-30 12:21:11 <clandmeter> ah
2020-10-30 12:21:17 <clandmeter> you are fetching master
2020-10-30 12:21:22 <mps> yes
2020-10-30 12:21:36 <clandmeter> this is for  your local testing?
2020-10-30 12:21:40 <mps> this patch 'must' be applied on master
2020-10-30 12:21:51 <mps> yes for local tests
2020-10-30 12:22:03 <mps> and ruby passed all tests with it
2020-10-30 12:22:41 <ikke> (where with the latest version in aports, it failed due to the crash patch that Ariadne added)
2020-10-30 12:23:26 <clandmeter> that does not make much sense for our gitlab issue
2020-10-30 12:23:28 <mps> now I'm trying to fix 'gc'
2020-10-30 12:23:42 <clandmeter> 3.11 to 3.12 has no chnage in musl
2020-10-30 12:23:46 <ikke> clandmeter: yes
2020-10-30 12:23:50 <mps> clandmeter: right
2020-10-30 12:23:54 <ikke> clandmeter: that's what I was referring to earlier
2020-10-30 12:23:56 <clandmeter> except few patching i think
2020-10-30 12:24:12 <mps> and I doubt this patch is relevant to 3.12
2020-10-30 12:24:42 <clandmeter> this patch fixes an issue introduced in musl which version?
2020-10-30 12:25:03 <ikke> 1.2.1
2020-10-30 12:26:29 <ikke> at least, the exacerbated it
2020-10-30 12:26:32 <ikke> I guess
2020-10-30 12:27:45 <clandmeter> right, so make no sense to spend time on this specific fix.
2020-10-30 12:29:52 <clandmeter> ikke: the deadlocks were in gitlab ruby? or another component?
2020-10-30 12:30:09 <clandmeter> gitaly i guess
2020-10-30 12:30:38 <ikke> gitaly-ssh
2020-10-30 12:31:11 <ncopa> i think we can test the patch in edge, and see if it solves the other known cases (like libvirtd etc) and does not introduce any other problems
2020-10-30 12:31:19 <ncopa> and then we can backport it to 3.12
2020-10-30 12:31:42 <clandmeter> https://gitlab.com/gitlab-org/gitaly/-/issues/3194
2020-10-30 12:32:11 <ncopa> oh..
2020-10-30 12:32:14 <ncopa> thats interesting
2020-10-30 12:32:40 <clandmeter> it sounds very familiar
2020-10-30 12:37:13 <clandmeter> ikke: there is a patch which we could try.
2020-10-30 12:37:34 <clandmeter> https://gitlab.com/gitlab-org/gitaly/-/merge_requests/2723/diffs?commit_id=891a604047dedcc82d552146062fbba6d2ea18e0
2020-10-30 12:38:16 <ikke> clandmeter: certainly worth a try
2020-10-30 12:39:27 <ikke> It does mention that they suspect clients which don't support sideband, which I assume their internal code would
2020-10-30 13:04:06 <Ariadne> to clarify i added the crash patch because it was hanging the builders
2020-10-30 13:04:08 <Ariadne> (:
2020-10-30 13:04:38 <Ariadne> i think a builder which fails immediately due to crashing is better than one which hangs forever and has to be restarted by hand
2020-10-30 13:04:50 <ikke> Ariadne: yes, certainly
2020-10-30 13:05:37 <ikke> safes me from killing those jobs every time :P
2020-10-30 13:08:23 <mps> Ariadne: ikke: I agree
2020-10-30 13:08:53 <Ariadne> sure i was just explaining the crash patch for clandmeter's benefit
2020-10-30 13:09:33 <Ariadne> however i am ok with backporting mt-fork into 3.13 release (and maybe 3.12)
2020-10-30 13:10:10 <Ariadne> i just haven't gotten around to testing heavily.  this week has been busy -- my grandparents are moving into an apartment in an assisted living community
2020-10-30 13:10:13 <ikke> is the patch backportable to musl<1.2?
2020-10-30 13:10:29 <mps> dalias told will make 1.2.2 soon, i.e. when we test it
2020-10-30 13:10:40 <Ariadne> no.  we would just upgrade 3.12 to 1.2.1 + mt-fork
2020-10-30 13:10:52 <ikke> ah, ok
2020-10-30 13:11:07 <mps> Ariadne: I' talking about edge
2020-10-30 13:11:23 <Ariadne> mps: yes, edge will become 3.13 soon
2020-10-30 13:11:25 <Ariadne> (:
2020-10-30 13:12:08 <mps> and not sure will simply add mt-fork patch for 1.2.1 will work correctly
2020-10-30 13:12:58 <mps> Ariadne: I (and I think ikke) testing musl master with mt-fork patch for few days
2020-10-30 13:13:38 <Ariadne> your thoughts, so far?
2020-10-30 13:14:38 <mps> rebuild rubu, zig, crystal and some smaller packages
2020-10-30 13:14:42 <ikke> mps: I have not really tested it
2020-10-30 13:14:43 <mps> no issues
2020-10-30 13:15:30 <mps> but with it found bug in 'gc' which I fixed temporarily
2020-10-30 13:16:23 <mps> even installed this updated/patched musl on bare metal aarch64, works fine
2020-10-30 13:17:01 <Ariadne> can't try on my honeycomb, but can try on my laptop
2020-10-30 13:17:41 <Ariadne> i think --disable-parallel-mark on boehm gc seems reasonable
2020-10-30 13:17:52 <Ariadne> i've been skeptical about that for a while
2020-10-30 13:17:54 <mps> ok, if you need patch I can post patch
2020-10-30 13:18:06 <Ariadne> nope, i upgraded based on the one you put on tpaste
2020-10-30 13:18:16 <mps> aha
2020-10-30 13:18:54 <mps> what about to make MR for gc, I have it ready
2020-10-30 13:19:07 <Ariadne> go for it
2020-10-30 13:19:10 <Ariadne> i'll merge it
2020-10-30 13:19:23 <mps> ok
2020-10-30 13:21:48 <Ariadne> let me find my yubikey
2020-10-30 13:24:02 <mps> nice, CI passed
2020-10-30 13:25:32 <mps> you are really fast ;)
2020-10-30 13:28:08 <Ariadne> just say yes
2020-10-30 13:28:12 <Ariadne> i don't need this delayed
2020-10-30 13:28:14 <Ariadne> who cares
2020-10-30 13:30:09 <mps> :)
2020-10-30 23:22:39 <Ariadne> ikke: it would be appreciated if i could get an s390x container.  the s390x VM i have outside of alpine's s390x setup is not very stable
2020-10-30 23:38:48 <ikke> I'll ask ncopa if I can arrange it
2020-10-31 10:07:02 <HRio> whats up with gitlab CI? 
2020-10-31 10:07:43 <HRio> lots of "Operation not permitted"
2020-10-31 11:39:26 <ikke> HRio: possibly due to musl changes
2020-10-31 11:39:43 <ikke> Need to check
2020-10-31 11:54:58 <ikke> https://tpaste.us/oPaR 
2020-10-31 15:11:10 <HRio> ikke: not much working :-)
2020-10-31 16:09:31 <ikke> :(
2020-10-31 16:10:39 <ikke> HRio: It's working with the latest version now
2020-10-31 16:10:49 <ikke> THe change that broke it has been reverted
2020-10-31 16:37:52 <HRio> ikke: thanks