| RikMills | Skia: could you wake MoM up please? | 07:15 |
|---|---|---|
| Skia | RikMills: done :-) | 10:02 |
| RikMills | Skia: ty :) | 10:02 |
| Skia | at one point I'll investigate to stop those hangings, but not today :D | 10:02 |
| === matttbe2 is now known as matttbe | ||
| ravikant_ | hey schopin, you may have some pointers here. there is this change added to busybox 1.37.0 https://git.busybox.net/busybox/commit/?id=40fa8eb998e27529fcd59830eb163b43658f9e81. It makes the build fail with `Support 64bit wide time types (TIME64) [Y/n/?] (NEW) aborted!`. I am looking at https://git.launchpad.net/ubuntu/+source/busybox/tree/scripts/kconfig/conf.c next. if I am missing something obvious, please let me know. | 11:44 |
| -ubottu:#ubuntu-devel- Commit 40fa8eb in busybox "build system: Make it possible to build with 64bit time_t" | 11:44 | |
| schopin | ravikant_: is this from a merge? | 11:45 |
| ravikant_ | yes | 11:45 |
| ravikant_ | some more information - debian buil does not have the same issue https://buildd.debian.org/status/logs.php?pkg=busybox&arch=amd64 | 11:45 |
| ravikant_ | `Restart config...` looks sus, hence I am looking at conf.c | 11:46 |
| ravikant_ | LP build log for referecne https://launchpad.net/~ravi-sharma/+archive/ubuntu/merge-lp2090999-plucky/+build/29375516 | 11:47 |
| schopin | ravikant_: I don't have any specific insight here, sorry. | 12:08 |
| schopin | ravikant_: fwiw building the vanilla Debian package on plucky seems to work fine. Assuming you're using g-u you can maybe isolate the specific bit of delta that triggers the build issue? | 12:11 |
| ravikant_ | schopin: thanks, that is very useful. | 12:12 |
| lvoytek | tsimonq2: Thanks for fixing up cacti and prometheus-mysqlrouter-exporter! That saved me a lot of time debugging. I'll add your prometheus-mysqlrouter-exporter fix to Debian too since I own the package there | 13:45 |
| === arif-ali_ is now known as arif-ali | ||
| bdrung | !dmb-ping | 16:00 |
| ubottu | bdmurray, bdrung, rbasak, schopin, teward, tsimonq2, utkarsh2102: DMB ping | 16:00 |
| schopin | bdrung: I'm confused. I thought it was in an hour? | 16:09 |
| bdrung | let me check | 16:10 |
| bdrung | the meeting is at 16:00 UTC - we are in winter time in UTC+1 | 16:11 |
| bdrung | schopin, "date -u" says Mo 9. Dez 16:12:04 UTC 2024 | 16:12 |
| schopin | OK | 16:12 |
| bdrung | !dmb-ping | 16:31 |
| ubottu | bdmurray, bdrung, rbasak, schopin, teward, tsimonq2, utkarsh2102: DMB ping | 16:31 |
| ahasenack | +1 maintenance: I'm working on backuppc and fence-agents, failures that resulted from /usr/bin/ping losing its privileges | 16:36 |
| schopin | ahasenack: this might be solved by the systemd merge. | 16:58 |
| ahasenack | schopin: because it has https://salsa.debian.org/systemd-team/systemd/-/commit/f2ddf70604920bc350c4f8aa7b895518d1808e33 ? | 16:58 |
| -ubottu:#ubuntu-devel- Commit f2ddf70 in systemd-team/systemd "sysctl: Add file trigger on /usr/lib/sysctl.d to restart systemd-sysctl" | 16:58 | |
| schopin | Yes. | 16:59 |
| ahasenack | in a local container (host: noble), that didn't work | 16:59 |
| ahasenack | I couldn't change net.ipv4.ping_group_range | 16:59 |
| schopin | ahasenack: what's your runtime? | 17:00 |
| ahasenack | 6.8.0-49-generic | 17:00 |
| schopin | I meant your container runtime :) | 17:00 |
| ahasenack | you mean this? lxd 5.21.2-084c8c8 31214 5.21/stable | 17:01 |
| schopin | Ugh, I think I've been playing with too many different scenarios and might have confused one terminal for the other. | 17:01 |
| schopin | Because it doesn't work for me either even though I could have sworn it used to. | 17:02 |
| ahasenack | so could I | 17:02 |
| ahasenack | $ sysctl net.ipv4.ping_group_range | 17:02 |
| ahasenack | net.ipv4.ping_group_range = 65534 65534 | 17:02 |
| ahasenack | I get that in a plucky fresh container, host noble | 17:02 |
| ahasenack | so let me try to change it | 17:02 |
| schopin | I can change it manually. | 17:02 |
| ahasenack | root@p:~# echo "0 2147483647" > /proc/sys/net/ipv4/ping_group_range | 17:03 |
| ahasenack | -bash: echo: write error: Invalid argument | 17:03 |
| ahasenack | is that the right syntax/values? | 17:03 |
| schopin | I just use `sysctl net.ipv4.ping_group_range="0 65534"` | 17:04 |
| ahasenack | but what is in the sysctl file, in /usr/lib/sysctl.d/50-blabla? | 17:04 |
| schopin | -net.ipv4.ping_group_range = 0 2147483647 | 17:05 |
| schopin | Which wasn't the value the runtime gave me, even after reboot. | 17:05 |
| ahasenack | that's what systemd will try to apply, and it will ignore errors | 17:06 |
| schopin | ooh, good point. | 17:06 |
| schopin | # sysctl net.ipv4.ping_group_range="0 2147483647" | 17:06 |
| schopin | sysctl: setting key "net.ipv4.ping_group_range": Invalid argument | 17:06 |
| ahasenack | don't you need a -w? | 17:07 |
| ahasenack | to write | 17:07 |
| ahasenack | but looks like it tried to write anyway | 17:07 |
| schopin | "he didn't know it was impossible, so he did it" | 17:07 |
| ahasenack | love that quote | 17:07 |
| schopin | Me too :) | 17:08 |
| schopin | The kernel doesn't seem to like the very big number. | 17:08 |
| ahasenack | the 65534 we get back on a fresh container seems to be the real gid, even though we are fetching it from inside the container | 17:08 |
| ahasenack | I think the large number there is just falling outside of the gid mapping range we give containers by default | 17:08 |
| ahasenack | this worked | 17:09 |
| ahasenack | ubuntu@p:~$ sudo sysctl net.ipv4.ping_group_range="0 1000" | 17:09 |
| ahasenack | net.ipv4.ping_group_range = 0 1000 | 17:09 |
| ahasenack | ubuntu@p:~$ ping 1 | 17:09 |
| ahasenack | PING 1 (0.0.0.1) 56(84) bytes of data. | 17:09 |
| ahasenack | I experience a similar problem when I use containers to test active directory joins | 17:09 |
| schopin | Yeah, I used 0 65534 and it was fine. | 17:09 |
| ahasenack | the gid/uid sssd hands out by default is large, and it falls outside the mapping we setup via /etc/sub{gid,uid} | 17:10 |
| ahasenack | but this all means that the new systemd won't be able to apply the new range from within containers | 17:10 |
| ahasenack | so the fix won't work | 17:10 |
| schopin | do we *need* the range to be that big? | 17:12 |
| schopin | If I understand correctly it's the container's gid that is taken into account by the kernel. | 17:13 |
| schopin | Otherwise we wouldn't be able to use ping with "0 1000" | 17:13 |
| * ahasenack tries 999 | 17:14 | |
| ahasenack | hm, worked for my 1000 uid (inside container) | 17:14 |
| schopin | wtf? | 17:14 |
| ahasenack | https://pastebin.ubuntu.com/p/3bWYT8pNqJ/ | 17:15 |
| ahasenack | so the host has a wide range, 0 - 2^31-1 | 17:15 |
| ahasenack | to the host, the container is uid 1001000 | 17:15 |
| ahasenack | (seen from ps aux from the host) | 17:16 |
| schopin | gotcha, but how come it didn't work at first then? | 17:16 |
| ahasenack | but the container has 0-999 | 17:16 |
| ahasenack | is this namespaced somehow? This setting? | 17:16 |
| ahasenack | and | 17:17 |
| ahasenack | ubuntu@p:~$ sudo getcap /usr/bin/ping | 17:17 |
| ahasenack | ubuntu@p:~$ | 17:17 |
| ahasenack | so definitely no capabilities | 17:17 |
| ahasenack | ok, this is weird | 17:17 |
| schopin | I have to run, would you mind dumping whatever you find at https://bugs.launchpad.net/ubuntu/+source/iputils/+bug/2089938 ? | 17:17 |
| -ubottu:#ubuntu-devel- Launchpad bug 2089938 in iputils (Ubuntu) "iputils 3:20240905-1 doesn't work for unprivileged users" [Undecided, New] | 17:17 | |
| ahasenack | "0 1"ah | 17:17 |
| ahasenack | ah | 17:17 |
| ahasenack | supplementary groups | 17:17 |
| ahasenack | uid=1000(ubuntu) gid=1000(ubuntu) groups=1000(ubuntu),4(adm),24(cdrom),27(sudo),30(dip),104(lxd) | 17:18 |
| ahasenack | \ | 17:18 |
| ahasenack | starting with 4 :) | 17:18 |
| tsimonq2 | lvoytek: You're welcome! Happy to help, and it saves a step for prometheus-mysqlrouter-exporter :) | 19:41 |
| tsimonq2 | lvoytek: The tl;dr is that cacti wasn't importing its initial database schemas correctly because of an error at the very top of its main SQL file, and prometheus-mysqlrouter-exporter needed a fixed config (should probably look at keeping that somewhat aligned with the default; I'll admit that I have no idea what the option I added does... but it works!) | 19:42 |
| tsimonq2 | \o/ mysql-8.4 is migrating (cc RikMills ) | 19:44 |
| RikMills | :D | 19:46 |
| lvoytek | :D | 19:48 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!