| onefang | So I often see this log entry - "unbound: [3932:0] info: generate keytag query _ta-4f66. NULL IN", any one have a clue what it means? What's the point of a commonly presented unintelligible log message? | 01:20 |
|---|---|---|
| onefang | It was almost once an hour this morning. lol | 01:20 |
| gnarface | i have no insight other than it seems to be coming from a unbound, a dns resolver i don't use, and it's class "info" which means you probably also have fairly verbose logging enabled for it | 01:23 |
| gnarface | so maybe it's not actually something you're even supposed to be seeing | 01:23 |
| onefang | Yep, unbound the DNS resolver, and default logging rules. | 01:23 |
| gnarface | is it always _ta-4f66, or does that number change? | 01:24 |
| onefang | Always _ta-4f66 this morning. | 01:26 |
| gnarface | it comes up in several search results, you aren't the only one asking about this | 01:27 |
| onefang | Same yesterday morning. | 01:27 |
| onefang | It was annoying enough this morning to trigger me to bitch about it BEFORE looking it up myself. | 01:28 |
| gnarface | seems to be related to DNSSEC, maybe a system time issue? | 01:28 |
| onefang | I run an NTP client. | 01:28 |
| gnarface | when you run ntpq, then type in "peers" at the prompt, make sure the entries have * in front of them | 01:29 |
| onefang | This is on my real hardware now, switched Deadulas from the test VM to be my daily driver now. | 01:29 |
| onefang | Well again, default setup for NTP, haven't looked at it yet. | 01:30 |
| gnarface | they changed from ntp to ntpsec, and there is a mistake in the default config file if you are only polling one server, it will just not use it | 01:30 |
| onefang | Actually I switched to chrony this time, but still default setup. | 01:30 |
| gnarface | you can tell because it will have + or nothing in front of the entries when you list the peers | 01:30 |
| gnarface | (this hit me too on upgrade to daedalus in several places) | 01:31 |
| onefang | Just double checked, ntp not installed, chrony is instead. | 01:31 |
| gnarface | hmm, dunno then | 01:32 |
| gnarface | eyeball check of the time looks accurate? for DNSSEC to fail it'd have to be off by like 5 minutes | 01:32 |
| onefang | Thu 25 Jul 2024 09:33:00 AEST | 01:33 |
| gnarface | maybe that's not even what was happening here, but it seemed to be an accepted suggestion in one of the search results... | 01:33 |
| onefang | Matches what it says on my phone. | 01:33 |
| onefang | My phone is Android and does use NTP. | 01:33 |
| gnarface | should be right | 01:34 |
| onefang | According to the timestamp I posted that at 09:33:09, so about right it took me 9 seconds to copy paste that. B-) | 01:34 |
| rwp | onefang, Here are the BIND docs on Trust Anchor telemetry, which is what you are seeing unbound logging: https://kb.isc.org/docs/aa-01528 | 02:24 |
| rwp | The corresponding unbound documentation is less than satisfying: https://unbound.docs.nlnetlabs.nl/en/latest/getting-started/configuration.html | 02:25 |
| rwp | The executive summary is that the log entry you cited "info: generate keytag query _ta-4f66. NULL IN" is a normal event using KSK-2017 (key ID 20326). I only see a couple of those a day logged in my log files. | 02:31 |
| onefang | Thanks gnarface and rwp. Good to know I can not only ignore them, but tell logcheck to ignore them as well. | 02:35 |
| * onefang moves onto whatever the next annoyance to fix will be. Ah the joys of upgrading by installing from scratch and going over all the configs again. Though I still compare them to the old ones, and sometimes just copy things. | 02:37 | |
| onefang | Today's annoyances I promised myself would be tmux, which ironically always does change their config in incompatible ways, except for this upgrade. So today I promised myself I'll work on the tmux annoyances that where always there. PROGRESS! lol | 02:42 |
| rustyaxe | ive not changed my tmux config file in 10 years, still works :P | 02:55 |
| onefang | Guess I was just unlucky and tried to use the config options that would change. | 02:56 |
| rustyaxe | Seems so :( | 02:56 |
| Xenguy | onefang, When I upgrade from Beowulf to Chimaera my tmux.conf broke, and I had to fix it | 03:19 |
| Xenguy | I have a sample file somewhere; let me know if that might be helpful | 03:19 |
| Xenguy | Their format changed IIRC | 03:20 |
| Xenguy | Or they stopped supporting their old format finally | 03:20 |
| onefang | This is an upgrade from Chimaera to Daedulas, the only tmux upgrade that didn't break on me. Beowulf to Chimaera broke on me to, but I fixed that long ago. | 03:21 |
| onefang | Right now my tmux problem is that after you tell it to split a pane by a certain percentage, then change the size of your terminal window, the percentage is lost and one pane holds onto what ever number of lines it had after the percentage split, the other just grows. I'd rather the percentages persisted. Alas "change the size of your terminal window" includes from the very beginning, when you haven't ac | 03:26 |
| onefang | tually atteched to the tmux session yet in a script. Then it assumes 80x24, and I never see the requested percentages at all. | 03:26 |
| onefang | Race condition I think. | 03:27 |
| Xenguy | Next time I'll ask you to specify the problem first : -P | 03:37 |
| onefang | I wasn't asking for help with that specific problem. | 03:38 |
| onefang | Got it half sorted, it is a race condition. If I adjust the timing of my commands, it works. But I don't want to fix it that way, there's reasons why the commands are in that order. | 03:39 |
| onefang | Ah sorted, the docs where not clear. | 03:41 |
| onefang | "tmux new-session -d -x - -y -" the '-' bit after the -x and -y tell it to use the current size of the terminal, not it's built in default size, or specifying a fixed width and height with -x and -y. The docs mentioned '-', but wasn't clear where to put it. | 03:43 |
| onefang | Naturally once you have attached to the tmux session, it'll pick the smallest terminal size amongst those attached to it, but by then the percentage size calculation has already converted into lines / colums based on tho original size. | 03:45 |
| adhoc | they each get a '-' ? | 03:45 |
| onefang | Yep. | 03:45 |
| onefang | You could probably have a '-' for one, and something else for the other. Didn't try that, not my use case. | 03:46 |
| rwp | I have had to chase tmux configuration changes but it was specific tmux commands which changed. | 04:37 |
| rwp | Though tmux paste has changed behavior generally and it used to work better and now the paste select preview is much worse. I should look again because I stopped using it. | 04:38 |
| rwp | It looks better now. It's now defaulting to the most recent copy. Which is good. I can verify that in a previous version (still have one running) it chooses the oldest copy by default. Which is terrible. | 04:42 |
| onefang | OK, think I have several annoyances solved now, in tmux and other things. Only way to be sure is to reboot though. So, BRB. | 05:54 |
| adhoc | rebooting solves all manner of ill's. | 06:27 |
| adhoc | except fixing your MBR | 06:27 |
| rwp | Should we be worried that onefang left to reboot and has not yet returned? | 06:43 |
| fluffywolf | probably. lol | 06:52 |
| onefang | Well got most of the way there, after much tweaking. | 06:55 |
| onefang | One remaining problem I think is just plain unsolvable, so I'll drop that. | 06:55 |
| onefang | One more tweak, one more BRB. | 07:48 |
| onefang | OK, good enough now. B-) | 08:19 |
| * onefang relaxes for the rest of the day. | 08:20 | |
| dvbst | hello, i tried the upgrade from devuan 4 to devuan 5 yesterday and now i cant log in due to the screen being frozen with the message "[ 84.814904] Bluetooth: hci0: Opcode 0x2037 failed: -22". even if i open up another virtual terminal, then i can see the login tty for few seconds, but then this message writes on it, clears the screen, and then its only that message again | 15:01 |
| rrq | have you tried "single user" boot? ... adding an S to the boot line | 15:12 |
| rrq | or else adding "init=/bin/sh" to the boot line? | 15:15 |
| rrq | or else boot upa live iso for forensics | 15:17 |
| dvbst | sorry, my internet went off, i cant see the messages | 15:18 |
| rrq | best option is to use a live iso | 15:20 |
| dvbst | ye but what do i do on it | 15:20 |
| rrq | figure out what happened on the prior boot up | 15:21 |
| dvbst | it recovers journal | 15:21 |
| dvbst | enters stage 1 | 15:21 |
| dvbst | starts udevd | 15:21 |
| dvbst | synthesizes initial hotplug events | 15:21 |
| dvbst | waits for /dev to be fully populated | 15:21 |
| dvbst | iwlwifi fails all over the screen | 15:21 |
| dvbst | bluetooth hci0 fails | 15:21 |
| dvbst | sets up keyboard | 15:21 |
| dvbst | starts boot logger | 15:21 |
| dvbst | cleans up temporary files | 15:21 |
| dvbst | loads kernel modules: lp, ppdev, parport_pc | 15:21 |
| dvbst | starts remaining crypto disks | 15:22 |
| dvbst | checks file systems | 15:22 |
| dvbst | complains that the current boot sector and its backup are different | 15:22 |
| dvbst | mounts local filesystems | 15:22 |
| dvbst | mounts swap | 15:22 |
| dvbst | cleans up temporary files | 15:22 |
| dvbst | starts apparmor something | 15:22 |
| dvbst | sets kernel variables: sysctl | 15:22 |
| dvbst | configures network interfaces | 15:22 |
| dvbst | does dhcprequest (this takes a while) | 15:22 |
| dvbst | cleans up temporary files | 15:22 |
| dvbst | sets up alsa | 15:22 |
| dvbst | runs X | 15:22 |
| dvbst | runit leaves stage 1 | 15:22 |
| dvbst | runit leaves stage 2 | 15:22 |
| dvbst | runsvchdir: default: current | 15:22 |
| dvbst | sets up console font and keymap | 15:22 |
| dvbst | starts uuidd and virtlogd | 15:22 |
| dvbst | fails to start avahi | 15:22 |
| dvbst | starts bluetoothd, connection manager, dundee, libvirtd, ofonod, saned (it complains that theres no avahi) | 15:22 |
| dvbst | screen clears | 15:22 |
| dvbst | then the login tty appears | 15:22 |
| dvbst | system writes to it: ok: run: dbus: (pid 2011) 0s | 15:22 |
| dvbst | same line again | 15:22 |
| dvbst | ok: run: elogind: (pid 2004) 0s | 15:22 |
| dvbst | then it repeats the things from avahi to ofonod, same order | 15:23 |
| dvbst | screen clears | 15:23 |
| dvbst | [ 84.814904] Bluetooth: hci0: Opcode 0x2037 failed: -22 | 15:23 |
| dvbst | this is what happens | 15:23 |
| rrq | ok. I don't know runit. I'm sure someone else might do though. | 15:24 |
| dvbst | when i was updating, it did say that avahi is now abandoned and people need to rewrite the scripts | 15:26 |
| dvbst | but i thought that i waited so long with the upgrade that it is tested | 15:26 |
| rrq | maybe you can disable most of the start up? how does runit pick which things to start/ | 15:29 |
| dvbst | iirc then there is stuff in /etc/sv/ and /var/service/ but ive never played around with it and i dont want to break more things | 15:32 |
| dvbst | i have the default runit + lxqt setup, never done any configuration to it, and i just followed the steps on https://www.devuan.org/os/documentation/install-guides/daedalus/upgrade-to-daedalus | 15:33 |
| rrq | ok.. you'll need to hang on until there's a runit wiz around | 15:35 |
| rrq | but in general the forensic is likely to be to disable almost everything and then add them back one by one until you run into the problem again | 15:35 |
| dvbst | i mean the only thing that seems to be stopping me from using my system normally is the bluetooth daemon | 15:36 |
| dvbst | and i dont even use bluetooth, so im fine with just removing it completely | 15:37 |
| dvbst | but youre right, lets wait for someone who knows what theyre doing | 15:37 |
| frew | Hello, I'm using antix, but devuan is pretty close, so can I ask for help please? | 16:16 |
| djph | frew: sure, you can ask for help about devuan in the devuan channel ... | 16:17 |
| frew | I have a quistion about sysvinit | 16:18 |
| frew | https://www.linuxquestions.org/questions/linux-newbie-8/long-dhcpdiscover-on-eth0-using-wlan0-4175739668/ | 16:18 |
| rwp | frew, Unfortunately if you are asking about networking and WiFi then the init is probably not really associated strongly. Different systems do network completely differently. I don't know how Artix/Arch does networking but I imagine it is doing things completely different. | 17:00 |
| rwp | And in Devuan there are several choices for handling WiFi and networking. | 17:00 |
| rwp | For regular networking the usual traditional choice is "ifupdown" which is the standard choice for wired networking. | 17:01 |
| rwp | For WiFi most unfortunately the, it pains me to say this, best (very painful), is NetworkManager. Which presents a mobile laptop GUI for WiFi. And it will handle wired networking. | 17:02 |
| rwp | But there is also "connman" too. Or one might do it manually. Or there are other possible ways too. | 17:02 |
| rwp | Also in Devuan/Debian there is a choice between "auto" which is a synchronous boot time setup and "allow-hotplug" which is an event driven on-event setup. | 17:03 |
| rwp | Not knowing anything about Artix/Arch I am not sure how any of this will apply there. | 17:04 |
| rwp | But if you install Devuan then we would have a common basis for helping you! :-) Good luck! | 17:04 |
| frew | antix is debian based, and it uses some devuan repos and sysvinit as a main init, that's why I'm asking here | 17:07 |
| rwp | I thought antix was arch based. No? Okay. Then what I said would apply. But I have a work meeting for the next hour now and must focus on it. | 17:09 |
| Xenguy | frew, For Wifi networking, have a look at this page: https://www.devuan.org/os/documentation/install-guides/daedalus/network-configuration.html | 17:11 |
| frew | rwp, you mixed up with aRtix) | 17:12 |
| Xenguy | I used the 'Using ifupdown' section for Wifi on my laptops... | 17:12 |
| Xenguy | Once it's set up properly it works great, and I just can't stand connman or NetworkManager | 17:13 |
| Xenguy | Now I don't have to deal with either of them ever again : -) | 17:13 |
| al1r4d | frew, sorry, i cant help you | 17:33 |
| al1r4d | ¯\_ (ツ) _/¯ | 17:34 |
| al1r4d | Well you can use devuan networking config and paste to your system | 17:34 |
| al1r4d | https://www.antixforum.com/forums/topic/antix-vs-devian/ | 17:35 |
| al1r4d | hmm | 17:35 |
| mason | rwp: I'd posit that ifupdown is Just Fine for wifi. | 19:47 |
| mason | rwp, frew: https://bpa.st/SP6Q | 19:49 |
| mason | In that example, "some network" will be connected in the case that both "some network" and "some other network" are available at the same time, based on the listed priority. | 19:50 |
| rwp | mason, I have used ifupdown for wifi and it is very good for a non-mobile wifi connection to an access point that is always the same. | 20:15 |
| rwp | For a mobile device it's tedious and a pain to have to manually scan for the SSIDs and then update the wpa_supplicant.conf file with the passphrase for them. Certainly possible. I ran my laptop that way for a year. But it's tedious. | 20:16 |
| rwp | I have myself since returned to using the GUI, meaning I am now running NetworkManager on my laptop for wifi. And I really, really, really dislike NetworkManager. | 20:17 |
| rwp | I should go back to using connman which is okay. It's just got this very unusual user interface model. | 20:17 |
| Xenguy | rwp, You don't have to wpa_gui takes care of the wpa_supplicant.conf management pretty seamlessly. | 20:19 |
| Xenguy | So it works just fine for laptops (which is all I use these days) and public wifi for example | 20:20 |
| rwp | wpa_gui? Is that new? I guess I will need to try it. | 20:21 |
| Xenguy | I found connman very intrusive, so very glad to have the ifupdown option | 20:21 |
| Xenguy | rwp, It's documented here: https://www.devuan.org/os/documentation/install-guides/daedalus/network-configuration.html | 20:22 |
| rwp | If wpa_gui handles the scanning and selection and passphrase part of the problem then that covers all of my tedious complaints about it otherwise. | 20:22 |
| Xenguy | See the first section of the document; wpa_gui is mentioned in one of the config examples | 20:22 |
| Xenguy | rwp, Yep, does the scanning, stores the passphrase etc. Works for me, and I use between a half a dozen and a dozen public wifi instances | 20:23 |
| rwp | This is all very good information. I'll take it out for a test drive later. I would be much happier if I were not using NetworkManager. | 20:24 |
| rwp | For about a year I was using a StarLabs StarLite laptop with only a manual wpa_supplicant configuration. And 95% of the time it worked excellently! Because 95% of the time I was routinely visiting the same WiFi SSIDs repeatedly. Once configured then it worked without any interaction. | 20:25 |
| rwp | But 5% of the time I would visit a new location and need to set up a new SSID and that was when the tedious nature appeared. | 20:25 |
| rwp | That StarLite laptop battery started swelling and popped the keyboard up where I needed to remove the battery from it for safety and stop using it. Unfortunately the battery is currently unavailable. | 20:26 |
| Xenguy | I feel exactly the same way, and I'm very happy that rrq came up with this method and documented it. I used to use the one that relied on Python 2, and it worked fine up to Beowulf, but disappeared in Chimaera. | 20:26 |
| rwp | I shopped for and returned to using a Thinkpad again. And there was much rejoicing because the keyboard on the StarLite was not to my liking. Truly awful I would say but I know someone else who liked it well enough. | 20:26 |
| rwp | I previously used wicd for python2 and really was quite happy with wicd. But... python2. I am really shocked and surprised that no python person has ported it to python3 yet. Makes me wonder what makes it so difficult. I assume because it must be write only code inside of it. But then I did say python so I repeat myself. :-) | 20:28 |
| rwp | I do really like how reliable and fast nfs2 reboots. So nice. | 20:31 |
| Xenguy | wicd, that's the one (I'd already forgotten the name ; -) | 20:36 |
| djph | man, i miss that one. | 20:36 |
| Xenguy | rwp, My /etc/network/interfaces is very simple: https://bpa.st/3J7A | 20:36 |
| Xenguy | That does everything I need to use the documentation/ifupdown method for public wifi | 20:37 |
| rwp | That must also be including wpa_gui too to fill in the ssid scanning and selection and passphrase stuff. | 20:38 |
| Xenguy | Yes, and /etc/wpa_supplicant/wpa_supplicant.conf looks like this (only the first 3 lines are needed, wpa_gui takes care of adding the wifi networks): https://bpa.st/EOUA | 20:40 |
| Xenguy | The only other requirement is that the user needs to be added to the 'netdev' group (for access to wpa_supplicant.conf I think) | 20:40 |
| Xenguy | Ideally the latter file should be chmod'd to 600 if it contains any private passwords | 20:41 |
| rwp | Thanks for the examples. I'll convert my Daedulas laptop over to it and drive it around that way. And there will be much rejoicing! | 20:43 |
| Xenguy | Enjoy the freedom from connman/NM : -) | 20:45 |
| frew | I didn't get the point, but isn't all that you need to do to connect to a public wifi is edit a little "interfaces" conf? why you should use wpa_gui? ps: asking for future | 20:51 |
| frew | and thanks for ifupdown mention, finally discover how network should be configured normally in linux | 20:53 |
| rwp | ifupdown has been the standard in Debian systems for decades. | 20:54 |
| frew | once, sysvinit was a main init | 20:55 |
| rwp | frew, So you take your laptop mobile to the library and connect to the library wifi. The poster on the wall says it is PublicLibrary345 or PubLib765 and the passphrase is "readbooks". That information must be entered into the wpa_supplicant.conf file by some method before wpa_supplicant will know that it should connect to that SSID with that passphrase. | 20:56 |
| rwp | Then you take your laptop mobile to your friends house. You ask your friend the name of the house WiFi. It is KeepOut3 and "Go Away Now". That information must be edited into the wpa_supplicant.conf file. | 20:57 |
| rwp | You take your laptop to a coffee shop and don't see anything up on the wall announcing it. Probably it will be obvious. You scan for access point SSIDs. You see one called "The Human Bean WiFi". The coffee shop name is The Human Bean. You make an assumption and look up the phone number. You enter that SSID and phone number and connect to it. | 20:59 |
| rwp | Using a GUI to scan and select and enter passphrases is much easier than running the commands to do so manually. Though you can certainly run the commands manually too. | 21:00 |
| frew | You can reboot) But ok, I get the point. "wpa_supplicant -c /etc/wpa_supplicant.conf -i wlan0 -B && dhclient wlan0" works but it's inconvenient obviosly | 21:02 |
| rwp | frew has left but starting wpa_supplicant and dhclient are the easy non-tedious parts. Which were not the parts of scanning, selecting, passphrasing, that I was calling out. | 21:40 |
| golinux | People come and go but do they talk of Michelangelo? | 21:47 |
| golinux | Sorry couldn't resist. :D | 21:47 |
| mason | Clearly you dared. | 21:48 |
| golinux | Hehehe . . . | 21:49 |
| golinux | A shout-out to T. S. Eliot | 21:51 |
| rrq | rwp: with Xenguy's setup, you let wpa_supplicant run dhclient, by virtue of the "iface default" block | 22:34 |
| rrq | .. and there's also a "wpa_cli" command/program to operato on the configuration from the command line (complemntary to wpa_gui) | 22:35 |
| rrq | (but I agree about using gui is helpful when roaming) | 22:37 |
| paculino | Has anyone made a gui to setup the wpa_supplicant + dhclient route? | 22:38 |
| gnarface | don't all the gui network tools do that, basically? | 22:38 |
| paculino | network manager didn't work that way for me | 22:39 |
| gnarface | i guess there aren't many choices, i think most people do use network-manager... i don't though. wicd was popular around here but it didn't make it into daedalus | 22:40 |
| rrq | wpa_gui is not very new but useful. possibly "man wpa_action" is a reasonable entry point | 22:40 |
| mason | paculino: If you're not stuck on it being a GUI, there's always "iwlist wlan0 scanning | grep ESSID" | 22:42 |
| mason | iwlist is from the "wireless-tools" package. | 22:43 |
| paculino | Oh, I was just curious. I set mine up last winter | 22:43 |
| gnarface | but frew was right, you can just put the wifi config into /etc/network/interfaces directly, you don't need any other crap. you might want to change the permissions on the interfaces file to not be globally readable after you've got your wifi password in it, but it should work fine with the regular ifupdown tools if you get the field names right... you just can't store multiple "location profiles" in there, you have to | 22:43 |
| gnarface | use comments and hand-editing | 22:43 |
| gnarface | but it definitely works | 22:44 |
| mason | gnarface: Gets more complex if you regularly connect to a few networks. For that including stuff in wpa_supplicant.conf eases the pain. Enables automatic selection. | 22:44 |
| gnarface | well, the wpa_supplicant config itself has completely redundant capacity in this regard, but you are not actually required to even touch it | 22:44 |
| rrq | with Xenguy's setup you can well have multiple location profiles | 22:44 |
| rrq | for me it was a matter of reading the collection of man pages | 22:45 |
| rrq | and trialling a bit | 22:45 |
| mason | Won't his throw timeouts if ethernet isn't plugged in? | 22:46 |
| rrq | maybe; I was referring to the wifi set up | 22:48 |
| rrq | which consists of a link level setup tied to the interface and one or more location setups, including a default one, for location profiles | 22:49 |
| rrq | most people would only have the "default" network level setup | 22:51 |
| rrq | s/have/need/ | 22:51 |
| gnarface | (i just comment out the ethernet lines when i'm not using them) | 23:00 |
| mason | gnarface: Maybe consider the mii-tool trick I use. I can't take credit for it, but it only tries to bring up ethernet if there's a link. | 23:04 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!