dan9er[m] | Did anyone see my message from last night? I heard Matrix->IRC was broken | 00:33 |
---|---|---|
dan9er[m] | Oh, ok these messages didn't relay to Matrix: http://reisenweber.net/irclogs/libera/_devuan/_devuan.2022-09-12.log.html#t2022-09-12T07:39:34 | 00:43 |
dan9er[m] | Apologies rwp | 00:44 |
rwp | dan9er[m], No worries. Everything happens in its own time on its own schedule. | 02:14 |
golinux | We have nothing to do with matrix | 02:32 |
dan9er[m] | Yo rwp I've looked at /etc/resolv.conf. There is a search line with `ht.home` on it, that matches the router's URL so I think thats good | 07:28 |
dan9er[m] | There's 7 nameserver entries, I won't post exact IPs however I can reverse DNS them (on the borked laptop) via mtr(8) so I can figure out what they are | 07:30 |
dan9er[m] | Quad9 IPv4 primary, Quad9 IPv4 secondary, Router (suspected, no hops), ISP IPv6 primary, ISP IPv6 secondary, Quad9 IPv6 primary, Quad9 IPv6 secondary | 07:33 |
dan9er[m] | s/Router/Router IPv6/ | 07:34 |
gnarface | do they all work? like, can you request a name lookup of a 3rd party domain through each of them directly? | 07:35 |
dan9er[m] | I mean normally running dig(8) works just fine | 07:37 |
dan9er[m] | `$ dig google.com` works fine | 07:37 |
gnarface | yea but it doesn't use them all at once | 07:37 |
gnarface | if just one is bad you'd have to test each of them to find it | 07:37 |
dan9er[m] | What does dig do differently? | 07:39 |
dan9er[m] | dig can query DNS just fine but not ping or any other app | 07:40 |
gnarface | you can't ping google.com? | 07:40 |
gnarface | you sure? | 07:41 |
gnarface | try it right now | 07:41 |
onefang | dig @9.9.9.11 google.com | 07:41 |
dan9er[m] | Not via that domain name. I can ping IPs (9.9.9.9, 8.8.8.8, 1.1.1.1, etc.) just fine | 07:41 |
dan9er[m] | aaand I just opened the manpage for dig, it uses resolv.conf by default | 07:42 |
onefang | Which is what the @ in my command is for, tell it to use that specific DNS server. | 07:42 |
gnarface | are you able to ping the ip that dig gives you for google.com? | 07:42 |
dan9er[m] | Yep | 07:43 |
onefang | 9.9.9.11 is the Quad9 one I use. | 07:43 |
gnarface | dig gives you a ipv4 address for google.com? | 07:44 |
dan9er[m] | Yes, ipv4 | 07:45 |
dan9er[m] | It's supposed to return ipv6? | 07:45 |
gnarface | no, just doing sanity checks | 07:45 |
gnarface | does ping show you the ip it's trying to use? | 07:45 |
gnarface | try "ping -n google.com" | 07:45 |
onefang | Try that dig command I suggested, but go through all the IPs of the name servers in resolv.conf. See if they all return proper results. | 07:46 |
dan9er[m] | gnarface: Nothing for 1min, then it gives up with `ping: google.com: Temporary failure in name resolution` | 07:48 |
dan9er[m] | Same as without `-n` | 07:49 |
gnarface | vaguely rings a bell about a bug a long time ago, not sure exactly what though | 07:49 |
onefang | dan9er[m]: Are you trying what I suggested? lol | 07:50 |
gnarface | yes, do this sanity check^ | 07:50 |
gnarface | though i'd expect intermittent failures if it wasn't all of them | 07:50 |
gnarface | i vaguely recall this may have happened once to someone with a partially-upgraded system | 07:50 |
gnarface | some sort of library version issue or something | 07:51 |
gnarface | normally everything will use resolv.conf and spec behavior is to try the first nameserver then failover to the next after a timeout, then repeat until its out of nameservers | 07:51 |
dan9er[m] | onefang: I just did, they all returned different IPs though I think that's because google is doing DNS load-balancing bs. Lemme try example.com instead | 07:51 |
gnarface | some stuff tries to load-balance by picking a nameserver off the list at random | 07:52 |
gnarface | but everything should be using /etc/resolv.conf as the primary source of nameservers unless /etc/nsswitch.conf says otherwise | 07:52 |
gnarface | but the vague memory in my head about a bug related to this i think involved some library version issue preventing ping from finding any resolvers | 07:53 |
gnarface | if you need a second opinion you can try traceroute | 07:53 |
dan9er[m] | I did the `dig @` thing with example.com, everyone agrees that it's 142.251.39.110 | 07:55 |
onefang | Funny, I get 93.184.216.34. | 07:56 |
dan9er[m] | Actually why is my router and ISP DNS servers in my resolv.conf? I thought I set everything to Quad9 | 07:56 |
dan9er[m] | Oh uh whoops read wrong number from screen | 07:56 |
onefang | Yeah if using Quad9, it's usually coz you DON'T want to use your ISP. | 07:57 |
dan9er[m] | s/142.251.39.110/93.184.216.34/ | 07:57 |
gnarface | if you're using your router for dhcp then it would also be giving you its own dns servers | 07:58 |
gnarface | the default behavior of all the dhcp clients is to clobber the resolv.conf file | 07:58 |
gnarface | (few plastic toy routers assigned by ISPs give you enough control over the dhcp service to override their dns servers, it's a good argument for running your own dhcp server) | 07:59 |
dan9er[m] | Yeah but Quad9 IPv6 is on the bottom, below my router and ISP | 07:59 |
onefang | I was offline when dan9er[m] and rwp originally discussed your problem, so that's all the advice I can give, since I dons't know what your actual problem is. | 07:59 |
dan9er[m] | onefang: http://reisenweber.net/irclogs/libera/_devuan/_devuan.2022-09-12.log.html#t2022-09-12T06:44:43 | 08:00 |
onefang | You are in good hands with gnarface anyway. B-) | 08:01 |
dan9er[m] | NetworkManager might just be being stupid | 08:01 |
gnarface | we've pretty much gone through all my ideas too though, i'm starting to wonder if it's a install issue | 08:01 |
gnarface | networkmanager is a struggle for me too, i try to avoid it | 08:02 |
dan9er[m] | Hm looking at nsswitch.conf, `hosts` is set to `files mdns4_minimal [NOTFOUND=return] dns` | 08:03 |
dan9er[m] | It had worked a while ago, maybe an update busted it or I removed a recommended/suggested package I shouldntve | 08:05 |
gnarface | hosts: files dns | 08:09 |
gnarface | try just setting it to this | 08:09 |
gnarface | no idea what that other crap is | 08:09 |
gnarface | i really wish i remember what that issue i ran into before was, but i think it may have only affected ping, so that might be something you can search by | 08:10 |
onefang | "files" means to use the /etc/hosts file. | 08:10 |
onefang | Though my memory of that might be wrotg. lol | 08:10 |
onefang | As wrotg as my typing. | 08:10 |
gnarface | yes, that's accurate, but i don't know why mdns anything would be on a linux system, and i have no idea what this does: [NOTFOUND=return] but i can only assume it means it gives up before trying dns | 08:11 |
onefang | That was my guess to. | 08:11 |
dan9er[m] | mDNS is multicast DNS | 08:17 |
gnarface | oh, not msdns | 08:17 |
gnarface | you using it for something? | 08:17 |
gnarface | removing that and the [NOTFOUND=return] thing would be a good deductive test | 08:18 |
gnarface | seeing anything on that line besides "files dns" is distinctly weird to me | 08:19 |
dan9er[m] | Aha, reading nsswitch.conf(5) mappage, `[NOTFOUND=return]` indeed makes it give up if not found. the default is `continue` | 08:20 |
dan9er[m] | So what the fuck put that there | 08:20 |
onefang | That's one reason why I use etckeeper, it helps to answer such questions. | 08:21 |
onefang | In case you don't know, etckeeper autamotically puts /etc into git and commits after every apt change. | 08:22 |
dan9er[m] | yeah def installing that after this | 08:22 |
onefang | So you have to manually commit anything you change yourself. And it shows you if something changed behind your back. | 08:22 |
dan9er[m] | Doing a `apt search mdns` shows libnss and avahi stuff installed | 08:25 |
gnarface | avahi is another one i generally avoid | 08:25 |
onefang | Me to. | 08:26 |
dan9er[m] | I'm gonna purge those and see if that unfucks my nsswitch | 08:26 |
onefang | But too many other things depend on it being installed, so I just don't run it's daemon. | 08:26 |
gnarface | in general avahi has also presented itself as a frequent culprit on the list of "changing stuff behind your back that shouldn't change" offenders | 08:26 |
onefang | Reminds me, I was surprised that nVidia stuff got updated the other day, I don't run any nVidia hardware. Chromium depends on the nvidia-settings libraries. Pffft | 08:27 |
gnarface | actually depends on them or just recommends them? | 08:27 |
onefang | Depends. | 08:28 |
gnarface | yuck | 08:28 |
dan9er[m] | Yep purging set `hosts` back to `files dns` | 08:32 |
dan9er[m] | Looking at APT log, it appears libnss-mdns package is the culprit | 08:33 |
onefang | That makes sense. | 08:34 |
gnarface | does ping work right now though? | 08:34 |
dan9er[m] | no... lemme try restarting | 08:34 |
gnarface | hmm, restarting shouldn't be necessary for this i don't think... | 08:35 |
gnarface | unless something else was wrong too, but restarting might not help either, then | 08:35 |
gnarface | is this a fully updated chimaera install? | 08:35 |
dan9er[m] | yep | 08:36 |
dan9er[m] | I should remind that if I tether through the phone, it works just fine | 08:37 |
dan9er[m] | s/the/my/ | 08:37 |
dan9er[m] | After restart, still not working | 08:37 |
dan9er[m] | what the shit | 08:37 |
gnarface | what if you try with traceroute instead? | 08:38 |
dan9er[m] | Same thing | 08:38 |
gnarface | maybe the router is blocking icmp? | 08:38 |
gnarface | outbound? | 08:38 |
gnarface | hmm, no it would get a namelookup still before failing then | 08:39 |
gnarface | hmmm.... | 08:39 |
gnarface | you think google is blocking your pings? how many domains did you try to ping? | 08:40 |
dan9er[m] | I'm now pinging example.com | 08:40 |
gnarface | oh, and it doesn't work either? | 08:40 |
gnarface | make sure you got all your updates | 08:41 |
gnarface | could be an issue with ping itself, like i said i remember running into something like that | 08:41 |
dan9er[m] | Can't ping suckless.org, example.com, duckduckgo.com, or matrix.org | 08:43 |
dan9er[m] | And now APT stopped being able to get DNS .-. | 08:43 |
dan9er[m] | i'll just edit sources with IPs | 08:44 |
gnarface | you don't have to do that | 08:46 |
gnarface | you can just override the domain's ip in /etc/hosts | 08:46 |
gnarface | temporarily | 08:47 |
gnarface | and you shouldn't have to restart the whole machine for it to take effect, though you might have to restart some programs | 08:47 |
gnarface | like firefox | 08:47 |
gnarface | if stuff can't even find the right ip by using the hosts file then you have something deeper wrong with the install for sure | 08:49 |
gnarface | missing or mismatched packages possibly | 08:49 |
dan9er[m] | hosts worked, and my system is fully up to date | 08:49 |
gnarface | hmm, so possibly a network issue | 08:50 |
gnarface | you using a firewall or something? | 08:50 |
gnarface | note that dns may exhibit only partial functionality if you can't reach the server with both tcp and udp | 08:52 |
gnarface | oh | 08:52 |
gnarface | something else just came to mind, might not be the issue but you might not be using the dhcp client you think you are | 08:53 |
gnarface | the kernel has a built-in dhcp client too now, and in some weird cases i remember it taking precedent over for the userspace one | 08:53 |
gnarface | and it could use something other than resolv.conf, i don't remember | 08:54 |
dan9er[m] | idk, didn't install any firewall packages | 08:54 |
dan9er[m] | I gotta go sleep | 09:14 |
Necrodiver | hello everyone, i have an issue that i cant seem to fix, so im on a desktop live iso of devuan just to back up files and things. | 09:22 |
Necrodiver | I tried to update refracta earlier, Im on Ceres. It was giving me an rsync error. I restarted my computer, and then it didnt load into the login screen, it dropped me to tty, and it didnt recognize anything. it said (none) login: | 09:24 |
Necrodiver | i tried do to starx and it didnt work | 09:24 |
Necrodiver | startx | 09:24 |
Necrodiver | I dont have the logs because i cant get back into my original installed os | 09:25 |
Necrodiver | i honestly dont want to have to do a full reinstall if i dont have to | 09:25 |
Necrodiver | https://dpaste.org/6pHBQ | 09:27 |
Necrodiver | heres the paste of what the error was originally | 09:28 |
Necrodiver | if i need to go into vim and change some settings that will fix it, i can find the file through the live session since the hdd is mounted | 09:39 |
hagbard | I'd start with using the live system do have a look at the logs from the original system. Without finding an actual error message somewhere, it won't be possible to figure out what's wrong. | 09:47 |
Necrodiver | would that be i nthe rsyslog file? | 09:49 |
Necrodiver | ive never had to look in these logs before so i have no idea where they are located | 09:52 |
hagbard | I think the problem of not beeing able to login is entirely unrelated to rsync. rsync is for syncing files over network, and is not related to login. | 09:55 |
hagbard | logs are in /var/log | 09:55 |
hagbard | I'd start with looking at /var/log/boot | 09:55 |
Necrodiver | okay ill look there hold on | 10:05 |
Necrodiver | https://dpaste.org/YcDNL | 10:08 |
Necrodiver | this was the last boot record since i have left my pc on for a few days | 10:10 |
Necrodiver | any idea as to what might be the issue? | 10:25 |
Necrodiver | i noticed one thing, in the sept 13th log, it doesnt load /dev/sdb1, | 10:31 |
Necrodiver | and sdb1 is where my home directory is installed, as i have a separate boot and home hard drives | 10:31 |
Necrodiver | https://dpaste.org/tszOL | 10:35 |
Necrodiver | heres the paste from september 4th | 10:36 |
unixbsd_ | hello | 10:36 |
unixbsd_ | is chimaera either : Bullseye (11;) or Bookworm (12;) ? | 10:36 |
Necrodiver | bullseye | 10:37 |
Necrodiver | bookworm is testing, aka daedelus | 10:37 |
unixbsd_ | thank you knindly very much | 10:38 |
Necrodiver | and just for reference later, Ceres is unstable | 10:38 |
unixbsd_ | is there atable of equivalence on wikipedia= | 10:38 |
unixbsd_ | ? | 10:38 |
unixbsd_ | and ascii ? | 10:39 |
Necrodiver | that much i dont know, i started using refracta at refracta 10, which was buster iirc | 10:40 |
onefang | https://www.devuan.org/os/releases has the table. | 10:40 |
unixbsd_ | thankm you !"!!! | 10:41 |
gorski | Hi guys. Can I have latest chimaera netinstaller iso and where from? I can not find it as rebuilt keyrings. Only old versions. Thanks. | 11:47 |
Necrodiver | why not just sudo dpkg -i keyring? Its not that hard | 12:03 |
drizzt | Hi guys ! | 12:17 |
drizzt | found a BIG CRITICAL upgrade issue in repositories, which reders systems unbootable | 12:18 |
drizzt | new lsb-base packeage is now "empty" | 12:18 |
drizzt | lsb-base_11.2_all.deb -> lsb-base_11.3_all.deb | 12:19 |
drizzt | lsb-base_11.3 does not provide /lib/lsb/init-functions anymore | 12:20 |
drizzt | but /lib/init/rc (and many many more init scripts) include it | 12:20 |
onefang | I guess this is in unstable or testing, coz stable (chimaera) has 11.1. | 12:21 |
drizzt | other init scripts have not been updated regarding to this update | 12:21 |
drizzt | onefang: yep | 12:21 |
drizzt | that's where dev happens :) | 12:21 |
drizzt | unstable | 12:21 |
drizzt | so please revert repository index to reference lsb-base_11.2_all.deb | 12:22 |
onefang | One of the other Devuan devs will have to look into that. | 12:23 |
rrq | it's a debian issue | 12:24 |
Necrodiver | so THATS why i cant boot my pc | 12:24 |
onefang | Ah someone mentioned that on the mailing list to. | 12:25 |
Necrodiver | here are my boot logs | 12:25 |
onefang | "It seems like Debian is moving the file to sysvinit-utils." | 12:25 |
Necrodiver | https://dpaste.org/tszOL | 12:25 |
Necrodiver | from sept 4th | 12:25 |
Necrodiver | https://dpaste.org/YcDNL | 12:25 |
Necrodiver | this is from today | 12:25 |
Necrodiver | the one thing i noticed when looking through them is that in the sept 4th log, it was reading /dev/sdb1 | 12:26 |
drizzt | yep, did updates one or two days ago and it was still lsb-base_11.2_all.deb | 12:26 |
Necrodiver | i updated today | 12:26 |
Necrodiver | and yes im on ceres | 12:26 |
drizzt | Necrodiver: you can use /lib/lsb dir from any other system if you have one at hand | 12:26 |
drizzt | that fixes the proble | 12:26 |
drizzt | problem | 12:26 |
drizzt | or force lsb-base_11.2_all.deb | 12:27 |
Necrodiver | so copy the lsb dir from a different pc and drop it there? | 12:27 |
drizzt | which should still be around | 12:27 |
drizzt | Necrodiver: yep | 12:27 |
Necrodiver | im on a live iso right now | 12:27 |
Necrodiver | just so i could back up my pc | 12:27 |
Necrodiver | like all the important stuff | 12:27 |
drizzt | did it, and works (at least for all my arm64 systems, but that's a -all package, should be arch independent | 12:28 |
Necrodiver | im on 64bit | 12:28 |
Necrodiver | so it should be okay | 12:28 |
drizzt | and it contains only two text files, so it is | 12:28 |
Necrodiver | okay so from the live boot | 12:28 |
drizzt | yup | 12:28 |
Necrodiver | is there a way i can force the old file? | 12:28 |
drizzt | yes | 12:29 |
Necrodiver | i can access the drive and everything so | 12:29 |
drizzt | using apt and pinning | 12:29 |
drizzt | pin the version to 11.2 for lsb-base | 12:29 |
Necrodiver | okay let me look in the lsb directory real quick | 12:29 |
drizzt | do not have the syntax at hand though | 12:30 |
Necrodiver | so ill have to look that up | 12:30 |
drizzt | lib/lsb disapeared during update | 12:30 |
Necrodiver | boy i was ready to give up and do a fresh reinstall | 12:30 |
drizzt | /lib/lsb/ dir | 12:30 |
Necrodiver | you saved me a lot of time | 12:30 |
drizzt | fell on it this morning on an embedded system I'm tunning | 12:31 |
drizzt | tested on a few others, and found the problem | 12:31 |
Necrodiver | hmmm | 12:32 |
Necrodiver | i went into lib, dont see an lsb anywhere | 12:32 |
Necrodiver | is it a hidden folder? | 12:32 |
drizzt | it even creates update problem for some packages | 12:32 |
Necrodiver | yeah i couldnt update exim4 | 12:33 |
Necrodiver | it was giving me an rsync error | 12:33 |
drizzt | Necrodiver: /lib/lsb dir got removed by update | 12:33 |
Necrodiver | Ooooooh | 12:33 |
Necrodiver | thats why | 12:33 |
Necrodiver | now i get it | 12:34 |
drizzt | you need to get it from lsb-base_11.2_all.deb or from the one on a system you did not upgrade yet | 12:34 |
Necrodiver | where would the lsb-base deb file be located? | 12:34 |
Necrodiver | ugh debians direct download pages are all set to 11.3 for sid | 12:37 |
Necrodiver | mother fuck | 12:37 |
drizzt | http://fr.deb.devuan.org/merged/pool/DEBIAN/main/l/lsb/lsb-base_11.2_all.deb | 12:38 |
drizzt | wget it | 12:38 |
drizzt | then dpkg -x lsb-base_11.2_all.deb | 12:39 |
drizzt | then dpkg -x lsb-base_11.2_all.deb . | 12:39 |
drizzt | (need a target dir) | 12:39 |
drizzt | or dpkg -i lsb-base_11.2_all.deb | 12:39 |
drizzt | might be better | 12:39 |
Necrodiver | but if im on a live boot | 12:39 |
Necrodiver | it would install it directly to the liveboot system right/ | 12:40 |
drizzt | then -x to extract content to curr dir | 12:40 |
drizzt | and copy (cp -a) or move to /lib of target system | 12:40 |
Necrodiver | so with -x after the file put the target dir and itll install directly there | 12:43 |
Necrodiver | okay extracted to the file into lib | 12:44 |
Necrodiver | lets restart my pc and see if this werks | 12:46 |
Necrodiver | if not, ill be back on liveboot | 12:46 |
Necrodiver | lol | 12:46 |
Necrodiver | so drizzt, it didnt work | 12:53 |
Necrodiver | it still dropped me to tty | 12:53 |
Necrodiver | and didnt recognize my home dir at all | 12:53 |
Necrodiver | i did a smart check on the drive, it passed | 12:54 |
Necrodiver | jsut in case | 12:54 |
Necrodiver | i installed the lsb-base directly to the lib folder too | 12:54 |
Necrodiver | just like you said | 12:54 |
Necrodiver | maybe theres something in the code of the init file there that needs to be changed? | 12:55 |
Necrodiver | i have my boot drive and home dir on separate drives so maybe theres something to that? | 12:55 |
drizzt | nope, have it so here too | 12:57 |
Necrodiver | hmmm | 12:57 |
Necrodiver | i wonder what happened then | 12:58 |
Necrodiver | i did exactly as you instructed | 12:58 |
Necrodiver | i tried to do startx to force start an x session in tty, said it couldnt connect to network | 12:58 |
drizzt | you must have /lib/lsb/init-functions and maybe /lib/lsb/init-functions.d/00-verbose | 12:59 |
Necrodiver | ahhhh | 13:00 |
drizzt | I do not have /lib/lsb/init-functions.d/00-verbose on each system (got it on those I fixed, and file is in lsb-base_11.2_all.deb | 13:00 |
Necrodiver | i see what happened | 13:00 |
drizzt | got to go | 13:00 |
Necrodiver | looks like i made a booboo lol /media/devuan/d743f327-2148-49aa-a384-aaea8459dcd6/lib/lib/lsb/ | 13:01 |
drizzt | I'll be back in about 2 or 3 hours, sorry | 13:01 |
djph | Necrodiver: oops. | 13:01 |
* Necrodiver smacks head | 13:02 | |
Necrodiver | i didnt realize it created a new /lib dir inside of /lib lol | 13:03 |
Necrodiver | good thing i checked the path | 13:03 |
djph | yep | 13:03 |
Necrodiver | okay just to double check | 13:04 |
Necrodiver | the lsb dir should be in /lib | 13:05 |
Necrodiver | the directories i have listed there are lib (the one i dont need because i screwed up), live and lsb | 13:05 |
Necrodiver | for l | 13:05 |
Necrodiver | and it should be the same directory as the /bridge-utils | 13:06 |
Necrodiver | dont wanna screw this up again | 13:06 |
Necrodiver | lol | 13:06 |
Necrodiver | well ill give it a try this time, hope it werks | 13:07 |
Necrodiver | lol | 13:07 |
Necrodiver | if not ill bb again on live boot | 13:07 |
Necrodiver | it worked | 13:10 |
Necrodiver | woohoo | 13:10 |
Necrodiver | somethings weird though, my volume icon has a no sound mark on it but the sound works | 13:11 |
Necrodiver | btw thanks drizzt, i know you arent here right now but i wanted to at least say thank you | 13:16 |
Necrodiver | now i just need to figure out why i have the volume at 100% but it says the sound is muted | 13:22 |
onefang | Better than the other way around. | 13:24 |
Necrodiver | true | 13:28 |
Necrodiver | i aint complaining its just weird | 13:28 |
drizzt | back | 15:50 |
drizzt | for the volume part ... I don't know | 15:50 |
Necrodiver | wb | 15:50 |
Necrodiver | yeah i dont know either, sound works so i dont care | 15:50 |
Necrodiver | lol | 15:50 |
drizzt | I have some urgent work which stacked up while I was away .. and I already lost more time on this than I had for today :( | 15:51 |
Necrodiver | damn | 15:51 |
Necrodiver | that sucks | 15:51 |
drizzt | f**k systemd | 15:51 |
Necrodiver | lol | 15:51 |
drizzt | and all it's side effects ... | 15:51 |
Necrodiver | work requires you to use that garbage? | 15:51 |
drizzt | "Ho, it's not used on my system, it can be safely removed for everyone ..." | 15:52 |
Necrodiver | ... | 15:52 |
drizzt | no, I do not use systemd | 15:52 |
drizzt | but this is a side effect of debian dropping sysvinit ... | 15:52 |
Necrodiver | debian outright dropped it? | 15:53 |
Necrodiver | wtf | 15:53 |
Necrodiver | i didnt know that | 15:53 |
drizzt | and it's dev being "mono-culuture" | 15:53 |
Necrodiver | so debian itself is going systemd only going forward? | 15:53 |
Necrodiver | that sucks | 15:53 |
djph | didn't they outright drop it a while back, hence devuan? | 15:53 |
drizzt | I do not even know if it's still possible to boot a debian without systemd | 15:54 |
Necrodiver | i thought the devuan thing was because of systemd implementation, not dropped sysvinit | 15:54 |
Necrodiver | and the way they implemented it | 15:54 |
djph | well, I mean there were a number of reasons. I'm probably mixing things up a tick | 15:54 |
drizzt | it was because of "single init" position of debian, and replacing single init "sysvinit" by "systemd" | 15:54 |
Necrodiver | ah | 15:55 |
Necrodiver | i thought debian was using multiple inits | 15:55 |
drizzt | but replacing init is a long term task, and I think they still have bits from place to place | 15:55 |
Necrodiver | im newer to linux only been using it like 3 or 4 years | 15:55 |
drizzt | this obviously is one | 15:55 |
Necrodiver | yeha thats lame | 15:56 |
Necrodiver | like, debian is one of the first distros. Debian should be the first to be all about open use of various init systems | 15:57 |
Necrodiver | they dont abide by the philosophy very well lol | 15:57 |
djph | or y'know simply not being another flavor or RedHat | 15:58 |
Necrodiver | true | 15:58 |
Necrodiver | lol | 15:58 |
Necrodiver | the problem is now that poettering is working for microshit, even MORE reason for "protest" distros like devuan | 15:59 |
djph | but we might be diverging from the channel's topic; given the "opinion" nature this seems to be taking (so, chanops might tell us to go to #devuan-offtopic) | 15:59 |
drizzt | it used to be "the universal OS" ... it is not anymore :( | 15:59 |
drizzt | yep | 16:00 |
Necrodiver | okay i joined off topic we can take the convo there | 16:00 |
drizzt | well I do not really have time for it :( | 16:00 |
Necrodiver | awww | 16:00 |
Necrodiver | well me and djph can | 16:00 |
Necrodiver | lol | 16:00 |
Necrodiver | im just gonna go to the convenience store and grab something to drink and a snack | 16:01 |
Necrodiver | ill bb in a few minutes | 16:01 |
drizzt | I'll stick here for some time in case someone needs help with this issssue, but I really should concentrate on my pile of work | 16:01 |
buZz | thats how we're all here ;) | 16:06 |
buZz | btw there's #devuan-offtopic for just social stuff | 16:06 |
dan9er[m] | gnarface: At this point I'm thinking of flashing an installer USB, nuking NetworkManager and then reinstalling from that USB as an APT source | 16:40 |
gnarface | i can't tell you if it'll work or not, but it's worth a try if you have a backup | 16:41 |
sxpert | good afternoon | 16:51 |
sxpert | I am trying to upgrade an old debian 9 to ascii | 16:52 |
sxpert | at some point the upgrade bombs in update-initramfs with "E: /usr/share/initramfs-tools/hooks/udev failed with return 1." | 16:53 |
sxpert | I did some digging and it fails at the line "copy_exec /sbin/udevd /sbin | 16:53 |
sxpert | I did some digging and it fails at the line "copy_exec /sbin/udevd /sbin" | 16:53 |
gnarface | this is probably covered somewhere, like in the release notes or something like that | 16:54 |
gnarface | i forget what the issue is but maybe if you upgrade udev to eudev first it'll help | 16:55 |
sxpert | it looks like /sbin/udevd is a symlink which copy_file doesn't grok | 16:58 |
sxpert | no, that comes up right when I upgrade to eudev | 16:59 |
gnarface | i don't think they should be symlinks | 17:02 |
gnarface | they're not symlinks in the eudev package i have on ceres anyway | 17:03 |
sxpert | hah | 17:03 |
sxpert | lrwxrwxrwx 1 root root 12 Apr 6 2018 udevadm -> /bin/udevadm | 17:03 |
sxpert | lrwxrwxrwx 1 root root 26 Mar 12 2016 udevd -> /lib/systemd/systemd-udevd | 17:03 |
sxpert | I have those 2 | 17:04 |
sxpert | the udevd one points to nowhere | 17:04 |
gnarface | oh, well that's definitely a problem | 17:05 |
sxpert | should this be a file or a directory with stuff inside ? | 17:06 |
gnarface | udevd should be a binary | 17:08 |
gnarface | the symlink to udevadm is correct | 17:08 |
sxpert | I restored the snapshot and I indeed have a binary | 17:09 |
* sxpert copies | 17:09 | |
sxpert | replacing the symlink with the original binary worked | 17:12 |
sxpert | lets try going forward now | 17:13 |
sxpert | so that's a case that should be checked | 17:22 |
AlexLikeRock | good morning | 17:37 |
Kitty | What do people recommend for wifi management on devuan? | 18:41 |
Kitty | wicd used to be my goto, but doesn't work any more | 18:41 |
Necrodiver | i stopped using wicd and moved to connman | 18:42 |
Kitty | that just works, thanks! | 19:07 |
fluffywolf | I use wicd. | 19:15 |
rwp | connman seems to be the natural heir apparent for wicd. | 19:24 |
Necrodiver | i remember when i switched to unstable and wicd stopped working, it was the first thing a friend recommended me | 19:24 |
Necrodiver | hasnt failed me | 19:24 |
buZz | wicd is python2 alas, and no dev wanted to port it | 19:27 |
fsmithred | I found connman to be a pain in the neck if there are two routers in the house. I switched to network-manager. | 19:37 |
drizzt | hostapd (router) and wpa-supplicant (clients), and a few lines in a shell scripts, works for all home needs | 19:44 |
drizzt | might sound "old geek", but works accros updates, changes, abandonware, ... | 19:45 |
drizzt | learning and understanding is always better than "simple" gui ... | 19:45 |
Necrodiver | true | 19:46 |
Kitty | true | 19:48 |
Kitty | this is for my work laptop. | 19:48 |
Kitty | so I also have a trade off of hours spent setting up shiny new laptop vs hour not worked | 19:48 |
drizzt | in most cases you do not switch between more than one or two wifi networks | 19:49 |
drizzt | gui tools are usefull if you switch network often | 19:49 |
drizzt | if you use a single one, or two, then maintaining the tools takes more time than learning and doing the configuration once | 19:50 |
drizzt | I stopped using gnome for this ... fed of having to reconfigure everything every 6 months because the devs did not think that my setup was worth supporting anymore ... every 6 months | 19:51 |
Kitty | I haven't used gnome in well over a decade | 19:52 |
drizzt | got a working and stable config of fvwm which I did not need to "redo" for about 12 years ... | 19:52 |
Kitty | I use evilwm | 19:53 |
Kitty | same config since at least 2004 | 19:53 |
drizzt | took me three days in the beginning, (but I did a lot of customizations) | 19:53 |
drizzt | learning "raw" tools is never a waste of time | 19:53 |
fluffywolf | I find wpa_supplicant to be utterly horrdendus to interface with directly, and would much rather use another program to do it for me. | 19:53 |
drizzt | the choice of the ****vm is up to you, but you get the idea :) | 19:54 |
fluffywolf | I only do it when I get a barebones install and need to get it on wifi to download more packages. | 19:54 |
Kitty | I got my new work laptop today | 19:55 |
drizzt | wpa-supplicant is mainly a config file and a few commands to "mount" the network | 19:55 |
Kitty | and took the list of packages from the old one, and just installed them all (yay for xargs) | 19:55 |
drizzt | Kitty ??? | 19:55 |
drizzt | why re-install ??? | 19:55 |
drizzt | Linux is not windows ... | 19:55 |
drizzt | I haven't installed a linux for years | 19:56 |
Kitty | it was the fastest way to bootstrap the new box | 19:56 |
Necrodiver | sure you coulda done it with a few command in terminal | 19:56 |
Kitty | oh I did do it in the terminal... | 19:56 |
fluffywolf | bbl, lunch over, back to the salt mine. | 19:56 |
drizzt | non, the fastest way is to take a copy of your rootfs on an USB key, which you should have ahndy anyway in case something goes wrong | 19:56 |
Necrodiver | speaking of which | 19:57 |
Necrodiver | just in case i should probably make a snapshot of my hdd | 19:57 |
Necrodiver | then put it on a usb drive | 19:57 |
Kitty | usb stick would mean asking CISO for permission to use an usb stick. and the nightmare issues of FDE | 19:57 |
drizzt | miam time | 19:57 |
___used | The xargs+package name list install method only works well if the list is ordered, deps 1st. Ime. | 21:46 |
golinux | How can I throttle the size of my .xsession-errors | 22:02 |
golinux | ? | 22:02 |
golinux | It is currently over 15 MB | 22:03 |
golinux | I forget it's there and it just balloons. | 22:03 |
phogg | have fewer errors? You can truncate the file any time you like. You could even logrotate it | 22:06 |
Necrodiver | what does the exim4-base package do? It says something for mail but i dont use thunderbird or mutt or anything so do i really need it? | 22:07 |
golinux | phogg: Mostly GTK errors from the theme | 22:08 |
peterrooney | Necrodiver: on most systems, exim delivers system-generated mail to root or appointed user account. | 22:15 |
Necrodiver | is it really a necessary package though? | 22:17 |
Necrodiver | because the way i read it on the debian package page, it isnt, but the way you describe it sounds like it is | 22:17 |
debdog | golinux: you could write a script which is run on shutdown/reboot to check its size and if it exceeds $VALUE delete it. | 22:21 |
debdog | and, thanks for reminding me, mine is at > 600 MB right now | 22:21 |
debdog | due to my setup I'll have to add that to my ~/.bash_profile | 22:23 |
phogg | golinux: if you don't care to look at it just truncate the file now and then, if you do care then logrotate it so the max size is manageable | 22:26 |
debdog | good idea! | 22:31 |
debdog | where does truncate snip the file, top or bottom? there does not seem to be an option for that. | 22:33 |
peterrooney | Necrodiver: root getting mail about system events via exim is, it turns out, entirely optional | 22:33 |
Necrodiver | oof lol | 22:33 |
peterrooney | it kinda makes sense, linux does loads of things (eg embedded) where mail would never ever get checked | 22:36 |
Necrodiver | so i could delete it or not, basically | 22:43 |
drizzt | exim is indeed optional | 22:43 |
drizzt | in fact, "all userspace is optional, save for a single process, usually called init" (even if you can call it anything else) | 22:44 |
Necrodiver | lol | 22:45 |
drizzt | I already built embedded systems with only a single usefull process (though I kept some userspace from busybox for easier debug/check) | 22:50 |
drizzt | (though this is out of devuan scope, which did not exist at this time) | 22:52 |
Necrodiver | ah | 22:53 |
dan9er[m] | erm gnarface you might want a load of this | 22:55 |
dan9er[m] | So we know `$ ping example.com` doesn't work | 22:55 |
dan9er[m] | I installed strace to get more output on that | 22:56 |
drizzt | what is "mandatory" in a distribution is only mandatory from the distribution's packagers' point of view. It's their job : build a set of packages so that it creates something usable from their point of view | 22:56 |
dan9er[m] | `$ strace -o trace.log ping example.com` gives `ping: socket: Operation not permitted` | 22:57 |
drizzt | which sometimes happen to be also some other people's point of view | 22:57 |
dan9er[m] | but `# strace -o trace.log ping example.com` works just fine! | 22:58 |
dan9er[m] | BUT `# ping example.com` doesn't!! | 22:58 |
dan9er[m] | wtaf | 22:58 |
dan9er[m] | `# ping example.com` gives temp. failure in name resoulution | 22:59 |
drizzt | dan9er[m]: strace adds some timings, but that seems weird | 22:59 |
dan9er[m] | My network issue just refuses to be debugged | 23:02 |
fsmithred | dan9er[m], user can't run ping anymore. I guess someone decided it's a security issue. | 23:04 |
fsmithred | run 'dpkg-reconfigure iputils-ping' | 23:04 |
fsmithred | install that package if you don't have it. | 23:04 |
dan9er[m] | I recall that I did do get/setcap to run ping as non-root | 23:06 |
dan9er[m] | But still, that was way before this issue | 23:06 |
dan9er[m] | And also it's not just ping that can't use DNS | 23:06 |
debdog | golinux, phogg: truncate removes the end of files. for logs which start at the top that's worthless. for now I stick with deleting it at a certain size. that way at least it (mainly) holds the messages. | 23:09 |
phogg | debdog: when I say truncate I mean '> ~/.xsession-errors' which throws away all data | 23:10 |
phogg | debdog: I did suggest logrotate as an option, which allows all manner of policies | 23:10 |
drizzt | dan9er[m]: your '#' didn't mean that you where running both commands as root ? | 23:10 |
dan9er[m] | fsmithred: Hm after running `# dpkg-reconfigure iputils-ping` nothing changed. It changed /bin/ping's cap from `cap_net_admin,cap_net_raw=ep` to just `cap_net_admin`. Still can ping as non-root to external IPs but DNS still borked | 23:11 |
dan9er[m] | drizzt: `$ sudo strace -o trace.log ping example.com` | 23:12 |
debdog | phogg: haha, my bad, I thought you actually meant the truncate command (I have no clue what '> ~/.xsession-errors' does. logrotate, will have a look, in case it's not too complicated... | 23:14 |
rwp | golinux, debdog, Every so often I do like the others here have mentioned and "> ~/.xsession-errors" to truncate mine to zero. | 23:14 |
fsmithred | that works like that? | 23:15 |
fsmithred | redirect nothing? | 23:15 |
rwp | The ">file" redirects the output of the command line to the file. Which creates the file if it does not exist. Truncates the file if it does exist. | 23:15 |
rwp | And if there is no command on the command line then the redirection is still done. | 23:15 |
drizzt | ls /proc/sys/net/ipv4/|grep icmp | 23:15 |
golinux | I have always just deleted it and it respawns | 23:15 |
dan9er[m] | Woah hello, I tried `$ sudo strace -o trace.log ping example.com` again and now that no longer works, getting temp. fail in name res. | 23:16 |
rwp | There was an obscure shell on True64 or some such that I haven't seen in 30 years that would produce an error. Some some of us pedants use ": > file" where : is the noop true command. | 23:16 |
dan9er[m] | Alright time to look in the trace | 23:16 |
drizzt | you have a set of files from kernel that control icmp capabilities ... though I do not know what you should have there or what to send in | 23:17 |
rwp | dan9er[m], Note that the ping command traditionally is setuid-root in order to work. But of late OS-people have decided to switch to Linux capabilities instead. | 23:17 |
drizzt | dan9er[m]: temp. ?? | 23:17 |
debdog | rwp: still don't understand entirely. where would I put that '> ~/.xsession-errors'? | 23:18 |
rwp | debdog, I just do that when I think about it from the command line. | 23:18 |
rwp | My ~/.xsession-errors file is 1.8MB at the moment. So I must have thought about it within a few weeks. Too soon to do it again. So I haven't automated it. | 23:19 |
debdog | ahh, so it can be done anytime | 23:19 |
dan9er[m] | drizzt `ping: example.com: Temporary failure in name resolution` | 23:20 |
fsmithred | wtf, I can ping example.com. I had no idea it was real. | 23:20 |
rwp | X had an open file handle to that file descriptor and the file handle is open until you exit X. Truncating the file will bring it to zero. New output from X or any of the children application programs of X will write to that file. | 23:20 |
rwp | I think X keeps the file open. Not sure. It might re-open it at times. I don't know. But truncating it to discard the contents is always safe. | 23:21 |
phogg | debdog: the advantage of logrotate is that anyone out there is probably already familiar with it. If you don't know it then it's less helpful. '> file' is shell code that, in effect, makes a file of size zero (even if the file already existed) | 23:21 |
rwp | Yes. example.com, example.net, and example.org are all real domains. Intended to be used for all documentation purposes. | 23:21 |
drizzt | what do you have in /etc/resolv.conf | 23:22 |
rwp | So annoying when people write something silly as an example and use ping mydomain.com and the owners of mydomain.com get the traffic. Use example.com instead! | 23:22 |
fsmithred | my resov.conf? | 23:23 |
drizzt | get IP from another way, and ping IP directly | 23:23 |
fsmithred | nameserver127.0.0.1 | 23:23 |
rwp | fsmithred, I think the resolv.conf question was for dan9er[m]. | 23:23 |
fsmithred | ah, ok | 23:23 |
debdog | phogg: thanks! slowly I begin to understand | 23:23 |
drizzt | fsmithred: no, dan9er[m] | 23:23 |
drizzt | sorry | 23:23 |
fsmithred | I'm using unbound. | 23:23 |
rwp | fsmithred is running a local caching nameserver. (as I am too, also unbound) | 23:24 |
drizzt | ping 93.184.216.34 | 23:24 |
drizzt | (IP of example.com) | 23:24 |
debdog | my first entry in .xsession-errors is of May 2015. which proofs I never remember. | 23:24 |
rwp | dan9er[m], I didn't read all of the scrollback but what is in your /etc/resolv.conf file? And what is "ls -l /bin/ping" (is it setuid?) or is it "getcap /bin/ping" (/bin/ping cap_net_raw=ep)? | 23:25 |
dan9er[m] | drizzt: generated by NetworkManager, `nameserver` entries as follows `9.9.9.9`, `149.112.112.112`, `2620:fe::fe`, `2620:fe::9` | 23:25 |
rwp | dan9er[m], You are using an external nameserver starting with the quad-9 one. Only the first three are ever used. But no errors if you have more. | 23:26 |
rwp | dan9er[m], Can you ping it? "ping -c3 9.9.9.9" ?? | 23:26 |
* rrq has heard that some ISP block port 53 traffic | 23:27 | |
rwp | rrq, That would be pure Evil(TM). But I have been on captured-portal networks before that did that. I usually tunnel around the blockage. | 23:28 |
dan9er[m] | rwp: `$ ls -l /bin/ping` gives `-rwxr-xr-x 1 root root 77432 Feb 2 2021 /bin/ping`; `# getcap /bin/ping` gives `cap_net_raw=ep`; and I can ping any external IP just fine | 23:29 |
dan9er[m] | And also dig works just fine | 23:29 |
rwp | dan9er[m], Then you are using the new style Linux kernel capabilities allowing cap_net_raw=ep instead of setuid for ping. | 23:29 |
rwp | But note that when you "sudo strace -o trace.log ping example.com" that uses sudo to become root, and then as root it does things as root instead of using the capabilities. So the behavior will be slightly different. | 23:30 |
rwp | "You are in a maze of twisty little passages all different." | 23:31 |
dan9er[m] | `sudo ping example.com` doesn't work either, same temp. fail in name res | 23:31 |
drizzt | beacome really root and test | 23:32 |
drizzt | sudo su - | 23:32 |
rwp | dan9er[m], Bizarre problem! | 23:32 |
drizzt | to get rid of all user rights possible problems | 23:32 |
dan9er[m] | drizzt: same shit | 23:33 |
drizzt | (I hate sudo, it's almost always badly used) | 23:33 |
rwp | I would try the "sudo -i" thing too. (sudo -i is effectively sudo su -) | 23:33 |
dan9er[m] | I have an strace of a `$ ping example.com` invocation that I can paste, if anyone is able to grok it | 23:33 |
drizzt | then, this israther a DNS problem | 23:33 |
rwp | dan9er[m], Never know until we try. Please send us a paste bin URL for it. | 23:33 |
dan9er[m] | One sec lemme get it off my problem laptop | 23:34 |
rwp | drizzt, Put me in the "I like sudo" camp. But I also use "su -" a lot too. They are different. Both are useful individually. | 23:34 |
rwp | dan9er[m], Oh! You have to sneakernet it. Sorry! It might not be useful. I was assuming you had it otherwise online. | 23:35 |
drizzt | rwp that's the point | 23:35 |
drizzt | sudo is useful for what it's designed to : give some rights on some commands to some users | 23:35 |
rwp | dan9er[m], Also I am still unclear... Can you ping by IP address directly? Is the problem only DNS host name lookup? Or is IP routing also broken? | 23:36 |
dan9er[m] | Well not sneakernet! Local IPs work. I have Syncthing (but I guess netcat would work too) | 23:36 |
rwp | drizzt, I might also say, sudo is useful to give full rights to some users. (Restricting commands (which I also do) has proven more tricky for people.) | 23:36 |
rwp | dan9er[m], So... LAN IPs work. Do WAN IPs work? If not then that sounds like a default route problem. | 23:37 |
rwp | LAN connectivity is always direct peer to peer across the LAN. But WAN IPs are routed by the router, usually using the "default" route. "ip route show" | 23:38 |
rwp | And depending upon things "ip -6 route show" for IPv6 networks. Which will be used preferentially if configured. And should be configured these days. | 23:39 |
dan9er[m] | rwp: https://hastebin.com/ivuvumuyal.strace | 23:39 |
dan9er[m] | Any IPs work, but not domain names | 23:41 |
rwp | dan9er[m], ping's libc resolver module code tried to connect to 9.9.9.9 (line 158) 149.112.112.11 (line 164) 2620:fe::fe (line 170) and all failed. | 23:41 |
dan9er[m] | And presumably anything that uses libc resolver | 23:42 |
dan9er[m] | Does dig(8) do something different? | 23:43 |
rwp | And everything pretty much uses the libc resolver module. | 23:43 |
rwp | dig might. I don't know. | 23:43 |
dan9er[m] | Any idea why that would fail? | 23:43 |
rwp | I am still suspecting something wrong at the network level below the DNS level. I think if the network worked then DNS would work. | 23:44 |
dan9er[m] | selinux? | 23:46 |
rwp | I am also despising of avahi and mdns and see mdns4_minimal in your /etc/nsswitch.conf file. I would be inclined to "apt-get purge libnss-mdns" double check /etc/nsswitch.conf and then try dns lookups again. | 23:47 |
dan9er[m] | Oh I purged that a while ago | 23:47 |
rwp | The libnss-mdns postinst script installs that "mdns4_minimal [NOTFOUND=return]" part and purging it will remove it. | 23:47 |
dan9er[m] | It's gone | 23:48 |
rwp | Good! I have had problems with it. And by design it does things that I don't think are good to do. For one it is trusting. And I have no trust. | 23:48 |
rwp | (I was going to say I have no trust in my soul. But I do trust people. Some people. But I don't trust everyone sight unseen.) | 23:49 |
dan9er[m] | dnsmasq-base is installed | 23:50 |
rwp | dnsmasq-base won't interact at this point. It's pulled in if you have containers anything. | 23:50 |
rwp | This is a real stumper. I would be inclined to download the refracta live boot image. Boot it. And see if the live image is okay or not. | 23:52 |
rwp | The live image should work out of the box okay. And if it does then you know something is definitely wrong in your current installation. | 23:53 |
dan9er[m] | I actually installed through the live image | 23:53 |
rwp | And if the live image fails to work then that is also useful information. Then it is something about your network. Your router. Or something in there. | 23:53 |
dan9er[m] | Cause this is a Surface Laptop 1 and the KB doesn't work without a patched kernel. Needed `toram` | 23:54 |
rwp | Then you have the image available to boot. Good. I would try that. Unless others have a better idea. Try the live boot and test it for network IP connectivity and DNS functionality. | 23:54 |
rwp | Sorry. KB is what in that context? I am unfamiliar with the Surface laptop. | 23:54 |
dan9er[m] | Keyboard | 23:55 |
dan9er[m] | And btw, dnsmasq-base is installed as a recommendation for network-manager. I don't have docker or the like | 23:55 |
rwp | Then I think you can safely purge dnsmasq-base. If something else needs it then it can pull it back in. Simplifying is always good. | 23:56 |
rwp | For example, libvirt kvm uses dnsmasq-base for a local NAT default network for the virtual machine network. | 23:57 |
rwp | Soo.... That keyboard problem is nasty. Does that mean you have to bootstrap yourself using only the touchscreen? Gack! | 23:58 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!