=== chris14_ is now known as chris14 === esembee_ is now known as esembee [09:53] Hello, what are the implications removing the "ubuntu-server" and "update-notifier-common" packages ? [09:54] when upgrading packages, i got that error: https://pastebin.com/raw/SCpTRcSa ; probably because some weird black magics were done by the person handling some python-related scripts [09:56] the best solution would probably to fix the python error, but that machine is all but "best-practices", so i feel like it's something that person won't want/be able to do [09:56] ... but otoh i'd like to keep doing security updates; which is broken now /o\ [09:58] (so my temp. fix was trying to remove the "update-notifier-common" package, but when doing that it also wants to remove the "ubuntu-server" one [09:59] i would try a "sudo apt install --reinstall python3-chardet" first [10:00] yes, that makes sense [10:01] but i fear it'll break what the person did to make it's script working [10:02] (but hey, is that my problem ? i'm not sure) ... ... [10:03] done ... /me logs out from the server [10:03] thanks for the hint [10:03] yw [10:05] Hey everyone! I noticed some weird image build failures with server images on focal. There seems to be a weird dep issue somewhat related to ubuntu-advantage-tools: "software-properties-gtk : Depends: ubuntu-advantage-desktop-daemon but it is not installable" [10:28] sil2100: do you have universe enabled? [10:29] No, our official images only depend on main, we never build with universe enabled [10:30] ubuntu-advantage-desktop-daemon seems to be in universe on focal, but on the other hand, I don't see software-properties-gtk depending on it [10:30] Guess I can promote it, as it's part of a main source package === cpaelzer_ is now known as cpaelzer [15:14] icey[m]: https://bugs.launchpad.net/bugs/1998636 [15:14] -ubottu:#ubuntu-server- Launchpad bug 1998636 in ceph (Ubuntu) "Unable to load libceph_snappy.so due to undefined symbol _ZTIN6snappy6SourceE in snappy 1.1.9" [Undecided, Confirmed] [15:14] that's fun [15:19] no kidding :-/ [15:39] oof [15:39] i am feeling empathy for anyone that ran across that one [15:56] at first glance I thought "wow, now you package ceph as snap, too?", but this sounds almost as bad [15:58] frickler: https://snapcraft.io/microceph :) [16:03] ravage: cool, thx for the pointer [19:51] congrats on the new upload rights, lvoytek ! \o/ [19:54] thanls! [19:55] thanks! === chris15 is now known as chris14 === rsalveti_ is now known as rsalveti [20:47] I am constantly pulling logs from a server to troubleshoot. On my mac, I copy to clipboard and have a shortcut to post a gist to github. Any reason not to set up the same thing on a server? Or is there a better way to do this rather than ... actually, I can output to a dir and then keep rsyncing it into our local shared drive. [20:47] That could work [20:54] why not just let syslog send them to where you need it? [20:57] foo: you might be interested in the pastebinit package/command [20:58] But yeah, I also don't really follow your problem. I would just write a "grab remote log and write it to a local file" one liner. [21:14] I use | nc termbin.com 9999 *so often* [21:15] putting (unredacted) server logs on a public pastebin might not always be a good idea though... [21:16] both from a security & from a privacy PoV [21:19] so true :) luckily most of what I put there is just source code === kostkon_ is now known as kostkon [22:53] why is it even after I turned off swap, perf is still saying the swapper kernel thread has 15-25% overhead [22:54] and turning vm.swapiness to either extreme, even before removing swap, had no effect [22:55] maybe I dont' understand what perf is telling me [23:05] 2. The official description of the process is that swapper is executed when no other tasks are running on the CPU.In other words, swapper means that the CPU did nothing and ran to rest [23:08] generally, turning off swap is a bad idea also... [23:20] mybalzitch: perhaps now swapper is busy discarding and reloading TXT pages from your executables and libraries? [23:20] how would I configure the 01-netcfg (netplan) to have a network port up but in listen mode (as the target of a switch port mirror/capture) [23:23] found an article from 2019 that uses a blank config (https://askubuntu.com/questions/1037276/bring-up-but-dont-assign-address-with-netplan ) is this still the most appropriate way? [23:24] baldpope: Om, [23:24] sigh [23:25] baldpope: I'm curious, does tcpdump or whatever need the interface up in order to set it to promiscuous mode? [23:25] sarnold, I believe it does [23:25] it's funny, I've never even thought to ask before. heh. [23:25] baldpope: anyway that seems like a reasonable enough answer, hehe [23:26] and by default, it brings it up with dhcp which might be OK, since there's not a DHCP server on that port, but felt wrong [23:26] sarnold: I let it sit for a few hours [23:28] sarnold, thanks - interface is up and looks good - have some more work to do, but looks good thus far [23:34] mybalzitch: I wonder if swapin.bt from bpftrace could show which processes are doing swapins, and perhaps help confirm which processes are having trouble.. granted, swapping tends to hurt everything.. https://manpages.ubuntu.com/manpages/focal/en/man8/swapin.bt.8.html === chris14_ is now known as chris14 [23:37] sarnold: oooh nice thank you. it seems stopping all my services, evicting everything from swap, then starting stuff again may have addressed my initial complaint. I look forward to running that when the issue comes up again. I was looking for a command like this. thanks. [23:38] mybalzitch: I was surprised it took me so long to find it. heh. I expected the iovisor tools to have it, perhaps as a feature in the biotop or similar. go figure.