Afdal | Can someone tell me why I can't seem to enable my swap partition... | 00:02 |
---|---|---|
Afdal | I don't even know why it's off, apparently | 00:02 |
n4dir | htop | 00:03 |
Afdal | I do do swapoff; swapon regularly to clean out my swap | 00:03 |
n4dir | sorry, i wanted to use htop in a terminal, it was not related to here | 00:03 |
Afdal | lol | 00:03 |
n4dir | you got an entry in /etc/fstab ? man swapon tells me it is swapon -a (for all), truth is i don't really know much about swap, someone will | 00:05 |
n4dir | missed what you wrote. So you did swapon, but it didn't activate it | 00:06 |
Afdal | oh lol | 00:06 |
Afdal | I was forgetting the -a | 00:06 |
Afdal | haha, all better | 00:06 |
Afdal | thanks c: | 00:06 |
rustyaxe | rwp: library dependancies. they saw x libs when built, so used em sorta thing. Totally appropriate, just be neat to setup a local vm that builds a devuan-lite for me projects ;) | 00:06 |
n4dir | more detailed infos might be useful | 00:11 |
rwp | rustyaxe, Which things? For example I know there is both vim-nox and emacs-nox which I use. htop in the terminal with no X. I use ssh no X. I use mtr-tiny which is no X. | 00:21 |
rwp | Afdal, I am inclined to suggest "lsblk --fs" to list a nice tree representation of the disks and partitions on your system with the UUIDs associated with them mapped out and other useful information. It's a nice utility. | 00:23 |
rustyaxe | rwp: i'd have to dig into the libraries it uses, but one of the audio or video codecs the pbx uses drags in the whole hell storm of X stuff. I just let it do the thing for now but 16gb rootfs msata ;) | 00:23 |
rwp | Perhaps with "lsblk -f | less +/swap" where it will highlight the swap parts. | 00:24 |
Afdal | yeah I know all about lsblk | 00:25 |
rwp | Hmm... pbx codecs. That's rude that it would pull in X libs. | 00:25 |
Afdal | I'm just a dummy and forgot to give swapon a parameter :) | 00:25 |
rwp | Ah... I see. Yep. That would do it. | 00:26 |
rwp | The system should mount all swap listed in /etc/fstab automatically at boot time though. Why are you flushing swap (swapoff-swapon)? I consider it an advantage that the system can offload some of my never used ram to disk to make the ram available for other things. | 00:27 |
rwp | I assume that along with that you are also doing "echo 3 > /proc/sys/vm/drop_caches" to flush the caches too? | 00:28 |
Afdal | Because I'm playin a lot of memory hungry games and using a modern web browser D:< | 00:31 |
rwp | So it's related by dd has two option flags that are useful for times when I am writing a 7.5GB live boot ISO to USB which would normally put all of that only used once data into the file system buffer cache uselessly. dd's oflag=sync,direct will write the data to the device O_SYNC synchronously, what that when the write is done it's done no need to wait longer for a flush can eject immediately, and with O_DIRECT which will bypass the file system buffer cach | 00:31 |
Afdal | no actually, explain caches to me... | 00:31 |
rwp | Afdal, https://www.kernel.org/doc/html/latest/admin-guide/sysctl/vm.html?highlight=drop_caches | 00:32 |
Afdal | can I get a short description | 00:33 |
Afdal | instead of linking an encyclopedia... | 00:33 |
rwp | It's the authoritative source of information on it. But let me sumarize. | 00:33 |
rwp | The kernel caches file system stuff into ram so that instead of needing to read the disk and wait for it to spin (ram is also faster than SSDs too) it can just refer to the ram cache of it. The file system I/O buffer cache is what makes file operations fast on Linux. | 00:34 |
Afdal | Doesn't seem like something which would use a lot of space | 00:35 |
Afdal | what kind of stuff does it cache exactly? | 00:35 |
Afdal | I hope it's not huge files... | 00:35 |
n4dir | fun fact: the how-to by Eric Raymong which was in the debian repos: "with the huge amount of RAM we have these days, sometimes even 64 MB, we hardly need swap anymore" | 00:35 |
Afdal | :( | 00:36 |
rwp | Basically the file system buffer cache is part way to working out of a ram disk or tmpfs because if one is doing something like "git clone foo && cd foo && make && make check ; cd .. && rm -rf foo" all of that might be so fast that it will all happen in ram and never touch physical storage at all. | 00:36 |
rwp | Things like continuous integration can be really very fast. As compared to other systems without a good file system buffer cache. Because it might all be in ram. | 00:37 |
rustyaxe | i havnet had swap on a machine in sooo long | 00:37 |
* rustyaxe stares at the poor 32gb ram peasant box in the corner. Not even it asks for swap :o | 00:38 | |
rwp | With a tmpfs the ram is backed up by swap. So if you have a /tmp as tmpfs and it is declared to be larger then it is possible to write enough files to /tmp that the system will need to use swap to handle all of the data of it. So tmpfs is good but not completely magic and can't make more ram than we have. | 00:38 |
n4dir | i haven't seen swap being used on 2 Gig machines. | 00:38 |
n4dir | but i don't really look too often | 00:38 |
rwp | Note that you are getting the IRC typing this in fast and furious off the top of my head summary and it's not really very rigorous and one should read the real docs for the details. | 00:39 |
rwp | Honestly most people don't really consume all of the resources of most of their machines these days because really 2GB is quite large and 32GB is still very large. | 00:40 |
Afdal | You people are insane | 00:40 |
Afdal | to think 2GB is large anymore | 00:40 |
Afdal | a modern disgusting web browser will gobble that up in an instant | 00:41 |
rwp | I want to hear you start counting the bytes off starting at one. Let's hear it! See how long it takes you. | 00:41 |
Afdal | Firefox uses 300MB when zero tabs are loaded!!! | 00:41 |
rwp | And remember rustyaxe and I were just talking about headless network servers just a moment ago with no need for X on them at all. | 00:41 |
rwp | Just to prove I am not a complete Luddite though I will recall when I was doing work on an HP990 with 512GB of RAM and 80 cpu cores. Could compile the entire Linux kernel in 26 seconds. | 00:43 |
rwp | Meanwhile I was using the Debian netinstall ISO yesterday on a VM system with 768MB of RAM configured and the installer popped up a dialog warning me that the system memory did not meet the minimum requirements! I increased the VM ram allocation to 1024MB (1GB) to avoid the dialog. | 00:45 |
rwp | After I rescued the system and finished using the netinstall boot I reduced the size back to something more reasonable to spare the ram for other VMs that need it more. | 00:46 |
fluffywolf | firefox uses more than 2GB of ram to start unless you have no extensions, one tab, and that tab is one of devuan's pages. | 00:46 |
rustyaxe | 2gb large? my box made of scrap parts from ewaste has 32gb | 00:46 |
Afdal | #killallwebdevs | 00:47 |
fluffywolf | last year I tried running firefox on a box with 192MB, and it failed pretty miserably. | 00:47 |
rwp | I use Firefox so I see it as my biggest piggy process but Chromium/Chrome and the others are just as piggy too. It's a terrible world. | 00:47 |
rustyaxe | total used free shared buff/cache available | 00:47 |
rustyaxe | Mem: 47673 4014 40417 754 4591 43658 | 00:47 |
fluffywolf | gimp is also a memory hog, but it has a legitimate excuse. | 00:47 |
rustyaxe | 10 year old laptop | 00:47 |
rustyaxe | how much even do new ones come with? | 00:47 |
fluffywolf | I have 16GB here. firefox got oomkilled last night... [8981056.437367] Out of memory: Killed process 17031 (firefox-esr) total-vm:8320888kB, anon-rss:1547992kB, file-rss:0kB, shmem-rss:52464kB, UID:1000 pgtables:13000kB oom_score_adj:0 | 00:48 |
rwp | This is OT but most of my machines have 16GB and none of mine are up to 32GB yet. Maybe on the next scraphound hunt. | 00:48 |
rustyaxe | Afdal: Its quite possible to make interact html5 applications that arent bloatasaurus. | 00:49 |
Afdal | Let's not though | 00:49 |
Afdal | the web isn't for "applications" | 00:49 |
Afdal | it's for loading documents for people to browse >:/ | 00:49 |
rustyaxe | use gopher if you want flat text | 00:49 |
fluffywolf | bluetooth headphones attempt #2... | 02:24 |
fluffywolf | yay, they work, and are actually the model I ordered! | 02:26 |
fluffywolf | they sound pretty decent. I'll need to listen to more music to give a proper review. | 02:28 |
fluffywolf | LDAC actually works now, given as they're the right model. heh. | 02:29 |
fluffywolf | so, next question... how the hell do I send firefox to the headphones? | 02:30 |
* fluffywolf is currently testing with https://www.youtube.com/watch?v=HlBni5HDsBY | 02:31 | |
fsmithred | fluffywolf, is sound coming out of the speakers when headphones are plugged in? | 02:32 |
fsmithred | and is it mini-jack or usb headphones? | 02:32 |
fluffywolf | <fluffywolf:#devuan> bluetooth headphones attempt #2... | 02:33 |
fluffywolf | :) | 02:33 |
fsmithred | ok, I have no idea | 02:33 |
fsmithred | if any sound is coming out of them, I'd expect all sound to come out | 02:34 |
fluffywolf | bluealsa sets them up as a separate alsa output device | 02:34 |
fluffywolf | for example, with mpv and friends, I can send audio to them with --audio-device=alsa/bluealsa | 02:35 |
fsmithred | it shows up a separate device with aplay -l ? | 02:35 |
fluffywolf | and firefox doesn't let me specify the output device, and changing default and reloading alsa will get annoying... | 02:35 |
fsmithred | yup. It will. | 02:36 |
fluffywolf | no, because aplay -l only lists things it thinks are physical cards. you have to do aplay -L to list it. | 02:36 |
fsmithred | maybe some asoundrc magic, but I don't know it. | 02:37 |
fsmithred | There's a post on the forum that shows how to use jack with firefox. Maybe something similar could get it to use the bluetooth | 02:39 |
fluffywolf | I also need to figure out an init script for bluealsa... | 02:40 |
fluffywolf | what is the long-term devuan plan for packages no longer including init scripts? | 02:40 |
fsmithred | hope they get added to orphan-sysvinit-scripts I guess | 02:41 |
fluffywolf | I'm tempted to write a proper script to use .service files, since sysd2v.sh does most of the heavy lifting | 02:43 |
fsmithred | go for it | 02:43 |
fluffywolf | I need to do a lot of RTFMing. no clue how init-d-script works... | 02:45 |
fluffywolf | also, right now I need to listen to a bunch of music to see how these headphones sound. :) | 02:46 |
fsmithred | :) | 02:46 |
fluffywolf | so far they sound good, but I might want a little treble boost | 02:46 |
fluffywolf | also, the debian bluealsa package seems to be built with debugging on, need to figure out how to turn it off... spams any terminal that loads the alsa lib quite a bit. | 02:47 |
fluffywolf | grr, bluetoothctl just crashed. it shouldn't do that. | 02:48 |
fluffywolf | googling seems to be suggesting pipewire to make firefox work; I've never used it. | 02:53 |
fsmithred | what does "hi" mean at the beginning of 'dpkg -l ' output? | 03:01 |
fsmithred | i means installed but what is h? | 03:02 |
onefang | It's just being friendly. | 03:05 |
debdog | H = Half-installed cannot find "h" | 03:14 |
debdog | "H = Half-installed", cannot find "h" | 03:14 |
debdog | there is a "h = Hold" for "Desired action:" | 03:14 |
debdog | man dpkg-query | 03:15 |
fsmithred | yeah, hold. Thanks. | 03:17 |
fluffywolf | the more I read about pipewire, the less I have any clue if it's useful or not. | 03:45 |
fluffywolf | I mean... https://bootlin.com/blog/an-introduction-to-pipewire/ is a simple introduction. | 03:46 |
rrq | fluffywolf: if you start firefox with like "apulse firefox" then firefox will play to the "default" pcm | 07:47 |
firefly_ | will it run on asus | 07:48 |
rrq | yes I would think so. apulse is an LD_PRELOAD | 07:50 |
gnu_srs1 | Hello, I'm trying to derive a list of packages causing problems due to usrmerge. Can I get some help from here? | 11:00 |
gnarface | gnu_srs1: yea, but you're gonna have to be patient, everyone is asleep right now | 11:05 |
u-amarsh04 | gnu_srs1 I had more of a problem with /usrmerge preventing or messing up booting | 12:21 |
u-amarsh04 | firmware being moved from /lib/firmware to /usr/lib/firmware and not installed into the initramfs by initramfs-tools | 12:22 |
u-amarsh04 | see this thread: https://dev1galaxy.org/viewtopic.php?id=6290 | 12:23 |
peterrooney | it's almost funny watching the pretzels arising from usrmerg on the debian-devel ml. | 12:29 |
JackFrost | That's actually not usrmerge, splitusr is what's causing the failure to boot there. If you *did* merge, /lib/firmware would have just been a symlink to /usr/lib/firmware thus picked up. | 12:32 |
firefly2 | i shall install devuan some day! | 12:37 |
djph | firefly2: why not today? :D | 12:48 |
gnarface | try a live cd at least | 12:55 |
djph | ^ | 13:37 |
fsmithred | I had to reinstall a bunch of packages because of various files that got lost when I ran usrmerge on ceres a few days ago. | 13:48 |
fsmithred | couldn't boot the 6.6.9 kernel either. Did that upgrade before installing usrmerge. | 13:50 |
fsmithred | after usrmerge, in several passes I reinstalled lvm2 dmsetup xorg xinit xauth and every installed package with xserver in its name, eudev elogind libpam-elogind policykit-1-gnome | 13:51 |
fsmithred | ^^^ from memory. Maybe something else, too. Maybe you will have something else. | 13:52 |
fsmithred | I can now boot into 6.6.9 kernel and run startx as user. Still couldn't get lightdm to run. | 13:52 |
fsmithred | On another VM I installed usrmerge in daedalus and then upgraded to ceres and had no problems. | 13:53 |
MovedMacBook | hello there... i am used to sharing my ethernet connection using wifi hotspot to my phone via networkmanager - but it doesnt seem to be working on the connection i setup , or by deleting a new connection and creating one.. even after a reboot or by creating a test connection as suggested in | 18:48 |
MovedMacBook | https://wiki.debian.org/WiFi/HowToUse#Troubleshooting_.26_Tips_for_NetworkManager or the top answer to this question - https://unix.stackexchange.com/questions/234552/ | 18:48 |
MovedMacBook | it was working all well until last week though.. not sure what caused this issue though | 18:49 |
MovedMacBook | or at least when i was on the kernel version - Linux devuan 6.1.0-16-amd64 | 18:50 |
MovedMacBook | but being on 6.1.0-17-amd64 seems to cause this issue .. probably because this is the only change that has happened in last 2 weeks | 18:50 |
gnarface | can you go back to the old kernel to check? | 18:53 |
MovedMacBook | i think i should be able to do that via grub.. | 18:54 |
MovedMacBook | if i remember it right that / some other old kernel version should be available | 18:54 |
MovedMacBook | let me give a reboot | 18:54 |
MovedMacBook | brb to update you all on status | 18:54 |
MovedMacBook2Dev | sadly it isnt working on 6.1.0-16 as well post reboot | 19:03 |
MovedMacBook2Dev | actually the problem is wifi hotspot gets started - but it is not delivering an IP to the clients - in this my case my android or my friends android.. which previously it did | 19:04 |
MovedMacBook2Dev | the hotspot is available to connect and authenticates to given psk - but android says it is obtaining IP address and waits for it.. and then disconnects | 19:05 |
MovedMacBook2Dev | after a timeout | 19:05 |
MovedMacBook2Dev | this was working in devuan - i remember setting up a hotspot connection via network manager UI as a connect automatically connection - so whenever i turn on wifi on macbook it turns on a wifi hotspot.. and i can connect to wifi the hotspot | 19:06 |
MovedMacBook2Dev | any directions to debug and fix this can be of much help | 19:07 |
gnarface | sounds like there is something wrong with the dhcp server | 19:09 |
gnarface | maybe it's either not running or starting in the wrong order | 19:09 |
gnarface | if you check the /var/log/daemon.log it should show some debugging info | 19:09 |
MovedMacBook2Dev | if i remember well - the p2p-dev-wlan0 in this config - https://dpaste.org/STNVW - will also be reported as connected whenever the connection works | 19:10 |
gnarface | try tailing the file while you try to connect, look for dhcp timeouts or something like that | 19:10 |
gnarface | do this both on the clients and the server | 19:10 |
MovedMacBook2Dev | which daemon.log should i tail? | 19:10 |
gnarface | the one called daemon.log | 19:11 |
MovedMacBook2Dev | because there is only these things that start with d | 19:11 |
MovedMacBook2Dev | $ tail -f /var/log/d | 19:11 |
MovedMacBook2Dev | dmesg dmesg.0 dmesg.1.gz dmesg.2.gz dmesg.3.gz dmesg.4.gz dpkg.log dpkg.log.1 dpkg.log.2.gz dpkg.log.3.gz | 19:11 |
gnarface | do you have rsyslog installed? | 19:11 |
MovedMacBook2Dev | no.. apt install rsyslog then? | 19:12 |
gnarface | yes | 19:12 |
MovedMacBook2Dev | sadness - rsyslog is already the newest version (8.2302.0-1devuan1). | 19:12 |
MovedMacBook2Dev | then how come i aint getting /var/log/daemon.log | 19:13 |
MovedMacBook2Dev | on tailing the file to connect to look for dhcp timeouts - i think i wont be able to tail things on phone coz its android and now the only way to look at server logs seems sadness | 19:14 |
cousin_luigi | MovedMacBook2Dev: Perhaps you have separate logs for the dhcp server? | 19:17 |
MovedMacBook2Dev | probably.. i am not sure.. where those logs are ..how can i find them? | 19:18 |
gnarface | should be somewhere in /var/log, but as far as i knew they'd all log to daemon.log by default... | 19:19 |
gnarface | you might have had to turn up log verbosity but not turn it on altogether | 19:20 |
MovedMacBook2Dev | actually i added a minor change on how wifi works post sleep/hibernate - | 19:20 |
MovedMacBook2Dev | could that be an issue - https://wiki.debian.org/InstallingDebianOn/Apple/MacBookPro/Early-2015-13-inch#WiFi_won.27t_work_after_hibernation | 19:20 |
MovedMacBook2Dev | i see | 19:20 |
MovedMacBook2Dev | is rsyslog a command that i can read man page of ? | 19:20 |
gnarface | you sure the file wasn't there? maybe it was there but you couldn't read it because you weren't root. | 19:21 |
gnarface | yea, you can read man rsyslog.conf, but like i said you shouldn't have had to change anything | 19:21 |
gnarface | and if you did have to change something, it would be more likely in the dhcp server's config than the syslog | 19:22 |
gnarface | is your dhcp server running? | 19:22 |
MovedMacBook2Dev | i swear these are the only logs at /var/log - https://dpaste.org/dCTQm | 19:23 |
MovedMacBook2Dev | i am not sure if dhcp server is running - because i am unable to find logs for it..and i can read man pages of rsyslog.conf .. but unable to see where that config file is.. to look into the config where it is currently sending the logs | 19:25 |
gnarface | ps aux |grep dhcp | 19:26 |
gnarface | you don't need logs to find out what programs are running | 19:26 |
MovedMacBook2Dev | $ ps aux | grep dhcp | 19:27 |
MovedMacBook2Dev | nobody 2497 0.0 0.0 7748 1880 ? S 23:30 0:00 /usr/sbin/dnsmasq --conf-file=/dev/null --no-hosts --keep-in-foreground --bind-interfaces --except-interface=lo --clear-on-reload --strict-order --listen-address=10.42.0.1 --dhcp-range=10.42.0.10,10.42.0.254,60m | 19:27 |
MovedMacBook2Dev | --dhcp-leasefile=/var/lib/NetworkManager/dnsmasq-wlan0.leases --pid-file=/run/nm-dnsmasq-wlan0.pid --conf-dir=/etc/NetworkManager/dnsmasq-shared.d | 19:27 |
MovedMacBook2Dev | yes.. this is the ip range that it used to deliver to phone - android | 19:27 |
gnarface | alright, so it looks like it's running | 19:27 |
gnarface | i didn't look at any of your pastes, is there a networkmanager log or nm log or anything like that? | 19:28 |
MovedMacBook2Dev | on client side - i just reboot my android device and also my friends android to connect to that test wifi - but both clients/phones timeout on "obtaining IP address" post reboot | 19:29 |
MovedMacBook2Dev | yes my 1st paste is nmcli log | 19:29 |
MovedMacBook2Dev | the 2nd is the content of /var/log directory | 19:29 |
MovedMacBook2Dev | https://dpaste.org/STNVW - nmcli log | 19:29 |
gnarface | if it doesn't tell you what you need to know, try increasing logging verbosity for both dnsmasq and networkmanager, if that's possible | 19:30 |
gnarface | and try tailing THAT log while you connect | 19:32 |
gnarface | you can watch the network traffic with wireshark or tcpdump too if you want | 19:33 |
gnarface | but there's only so many things that can be going wrong here | 19:33 |
MovedMacBook2Dev | i can see some logs coming up in /var/log/syslog whenever i connect my phone to wifi hotspot | 19:38 |
rwp | Everything except security goes into the /var/log/syslog in addition to the other supplemental log files. I would look in /var/log/syslog and figure out the missing daemon.log later. Though that daemon.log is missing is perhaps a clue that we are actually on a different system or something. | 19:38 |
MovedMacBook2Dev | now to increase log level on nmcli | 19:38 |
MovedMacBook2Dev | but this is a devuan system - daedalus | 19:38 |
gnarface | did you install from live cd? | 19:39 |
gnarface | migrate from another system? | 19:39 |
MovedMacBook2Dev | rwp - you mean to say i customized it ? - i installed via netinst - pend drive image | 19:39 |
gnarface | if syslog is showing the relevant info, that's what matters | 19:40 |
gnarface | the missing daemon.log thing is suspicious though... | 19:40 |
MovedMacBook2Dev | if i remember the names right - @fsmithred helped me install via the installer - because it was my first time using that CLI based installer | 19:41 |
rwp | I never look anywhere but /var/log/syslog and /var/log/auth.log on Debian-like systems. (It's /var/log/messages on other systems.) | 19:41 |
rwp | Hmm... My Unstable system stopped logging to /var/log/daemon.log on Oct 30 2022 and has not logged there since. | 19:41 |
MovedMacBook2Dev | ok.. nothing on auth.log when my phone is trying to connect to hotspot | 19:42 |
rwp | On Oct 30 2022 my Unstable system upgraded from rsyslog 8.2210.0-2 to 8.2210.0-3 and at that time the rsyslog.conf was changed in the upstream packaging to remove logging to the daemon.log file. | 19:43 |
gnarface | hmm, maybe it's not a thing anymore then | 19:43 |
MovedMacBook2Dev | i do feel so | 19:43 |
gnarface | so does all that go to syslog now? | 19:44 |
rwp | Looks like daemon.log is no longer a thing. OldStable Chimaera has it but Stable Daedalus does not. | 19:45 |
rwp | Everything has always gone to syslog. Except for security things that goes to auth.log. | 19:45 |
MovedMacBook2Dev | sadly nmcli log level increase to debug level doesnt show us what is needed | 19:47 |
rwp | Other Linux and *BSD systems have logged to /var/log/messages. Other Unix systems such as HP-UX call it /var/adm/syslog/syslog.log. It's always been annoyingly inconsistent among different systems. | 19:47 |
MovedMacBook2Dev | these are the logs - https://dpaste.org/GOcnG | 19:47 |
gnarface | daemon.*-/var/log/daemon.log | 19:47 |
gnarface | that line, i guess | 19:47 |
MovedMacBook2Dev | interesting deviation on the derivates | 19:48 |
MovedMacBook2Dev | derivatives" | 19:48 |
rwp | If one installs a new Daedalus system that line won't be in the new file. | 19:48 |
MovedMacBook2Dev | which line? | 19:49 |
rwp | If one upgrades from an older system and tells dpkg to keep the old file then it will. I always tell dpkg to install the new version of conffiles and then merge as needed immediately following. | 19:49 |
MovedMacBook2Dev | i did install a new daedalus system about 2-3 weeks ago | 19:49 |
MovedMacBook2Dev | oh the upgrade pathway .. | 19:50 |
rwp | The line that gnarface just posted "daemon.* -/var/log/daemon.log". That has been removed from the upstream packaging for rsyslog in Daedalus. | 19:50 |
rwp | We have traditionally been able to upgrade systems. That used to work. Most of us want it to continue to work. But there are forces acting in opposition. | 19:50 |
MovedMacBook2Dev | that is sadness | 19:50 |
rwp | Much sadness. Have sad life. Probably have sad death. But at least there is symmetry. --Zathrus, Babylon 5 | 19:51 |
rwp | That was a great show. | 19:51 |
rwp | Anyway... I have never had anything but pain from NetworkManager. This feels just typical to me. | 19:52 |
rwp | I focused on the /var/log/daemon.log problem. Can we re-state the networking problem that started this? I got lost reading about it. | 19:53 |
MovedMacBook2Dev | oh.. | 19:53 |
MovedMacBook2Dev | actually this is my first with networkmanager / nmcli | 19:53 |
MovedMacBook2Dev | it always worked OOB for me.. | 19:53 |
rwp | People do use it. It remains very popular. But only because there is little competition. The main alternative is "connman" connection manager. It works. It is a little odd. | 19:54 |
MovedMacBook2Dev | even before 12 years ago on ubuntu about 4-6 ago on fedora .. and until last week or last kernel update on daedalus | 19:54 |
rwp | As I read things you were using your Devuan system running NetworkManager to provide a WiFi Access Point to which you could connect your phone. And now the Devuan WiFi AP is not working? Is that the current problem? | 19:55 |
MovedMacBook2Dev | also even debian wiki praises it - https://wiki.debian.org/WiFi/HowToUse#Troubleshooting_.26_Tips_for_NetworkManager | 19:55 |
MovedMacBook2Dev | true my lord - rwp | 19:55 |
MovedMacBook2Dev | my client - in this case - android phone - doesnt get a successful IP delivered post successful identification and authentication to the available wifi AP | 19:56 |
rwp | The Debian Project also praises systemd too but here we are. | 19:56 |
MovedMacBook2Dev | in this case test AP via the command - nmcli dev wifi hotspot ifname wlp4s0 ssid test password "test1234" - as suggested in debian wiki and also on stackexchage - https://unix.stackexchange.com/questions/234552/create-wireless-access-point-and-share-internet-connection-with-nmcli/384513#384513 | 19:57 |
MovedMacBook2Dev | hmmm.. thats true | 19:57 |
rwp | For debugging this I always think it is good to get a second client side opinion. Can any other device be a client and associate on the WiFi AP? Because, for example, it would be a shame to do all of this debugging of the server only to find that it was a client problem on the Android phone. | 19:58 |
MovedMacBook2Dev | i have another android phone.. which also timeouts on "obtaining IP address" | 19:58 |
rwp | Good double check. That's a good double check of the problem. | 19:59 |
MovedMacBook2Dev | i reboot both the phones still i get the same thing | 19:59 |
rwp | On the client side for a WiFi Access Point there are three component parts of an AP that all three need to work. 1) WiFi AP WAP/WAP2 association. 2) DHCP IP Address and Router assignment. 3) DNS. | 20:00 |
rwp | If any of those three things (and there are minor possibilities such as ntp server and some other things too) fail then the client will not be fully functioning. | 20:00 |
rwp | s/client side/server side/. On the server side for a WiFi access point those things must work. My bad. | 20:01 |
rwp | Can you tell what is working and what is not working from the logs and such? | 20:01 |
MovedMacBook2Dev | hmm... on the client side - android - i setup private DNS - an adblock Dns DoT server.. | 20:01 |
MovedMacBook2Dev | so i removed that.. even then it is obtaining IP timeout | 20:02 |
rwp | So the WiFi AP association is failing? Right at the earliest part of things? | 20:02 |
MovedMacBook2Dev | previously when my AP worked - my dns issue was never an issue.. it was workning well | 20:02 |
MovedMacBook2Dev | true.. it the 2nd step | 20:03 |
rwp | I just know that people will tell me that "the Internet is down" and I will find they have a misconfigured DNS server and not able to get any DNS and the Internet connection is fully functioning. Happens so often that I always check that first. | 20:03 |
MovedMacBook2Dev | it is the 2nd step.. | 20:03 |
rwp | 2nd step. Good actually. I think that is easier to debug. | 20:03 |
MovedMacBook2Dev | let me get another machine to check the client logs .. its a fedora laptop though | 20:04 |
MovedMacBook2Dev | sounds kewl to debug and faster.. am actually trying to get my head around this for about 36 hours now | 20:05 |
rwp | Earlier gnarface was examining the dnsmasq daemon running on your machine. That's when things went into the daemon.log and got snarled up. Let's get back to gnarface's questions there. | 20:05 |
MovedMacBook2Dev | sure.. | 20:06 |
MovedMacBook2Dev | when i increased the nmcli logging level.. nothing was observable on syslogs..now onto dnsmasq - b4 getting on to it.. should i decrease nmcli back to INFO level? | 20:06 |
rwp | I jumped in on the daemon.log and rsyslog configuration because I knew about that part. But I want to back away slowly now because I don't use NetworkManager myself and don't know much about the details of it. | 20:06 |
rwp | I do run my own software hostapd access points NOT using NetworkManager though and can talk about that part in general. But since you are using NM that info won't directly apply. | 20:07 |
MovedMacBook2Dev | i can use conman if that works - rather than debugging this.. coz its kinda late in here on the asian side | 20:07 |
MovedMacBook2Dev | because apt install conman is way faster and cheaper - i guess | 20:07 |
rwp | I am not using connman for other than a client. I don't know how it is configured to act as an Access Point. | 20:08 |
gnarface | MovedMacBook2Dev: you were looking for dhcp errors, i don't know if it'll matter about networkmanager's setting if nothing changed | 20:08 |
rwp | If you knew when things stopped working then another useful log file is /var/log/dpkg.log which will log package upgrades. Maybe find a package that changed at that time? | 20:09 |
MovedMacBook2Dev | i see.. so even when conman plays it part - the dhcp might still does not participate - is this what you are trying to conver gnarface? | 20:09 |
MovedMacBook2Dev | sadly nothing on dpkg.log - coz # wc -l /var/log/dpkg.log | 20:10 |
MovedMacBook2Dev | 0 /var/log/dpkg.log | 20:10 |
rwp | I think you meant to say s/conman/network-manager/ there? You are using network-manager and it is starting dnsmasq, right? | 20:10 |
MovedMacBook2Dev | i mean .. i dont have conman installed.. | 20:11 |
gnarface | well, usually with isc-dhcp-server, or dhcpcd, it would actually say why the dhcp handshake fails | 20:11 |
MovedMacBook2Dev | i can install it and try it out | 20:11 |
rwp | It would have been logrotated to one of the other /var/log/dpkg.log* files such as /var/log/dpkg.1 or dpkg.2.gz or older. | 20:11 |
gnarface | presumably dnsmasq will give similar info | 20:11 |
gnarface | and if not, consider changing it | 20:11 |
rwp | I agree. I am thinking that about dnsmasq too. | 20:11 |
gnarface | it might just need a timeout increased | 20:12 |
gnarface | or there might be some other error | 20:12 |
gnarface | going back to an older version of dnsmasq, if it recently changed too, that might also give some evidence... | 20:12 |
rwp | MovedMacBook2Dev, I know by what you said that you are thinking you can just swap dnsmasq out and install isc-dhcp-server and things like they were modular and if it were you doing this yourself then you could but NetworkManager does not work like that. With NM you can't just swap X for Y. NM wants to do things a certain way and you have to go along with it. | 20:13 |
rwp | People who use dnsmasq love it. Primarily because the entire configuration can be provided on the command line to it without a config file at all. The program works. But that doesn't mean something might not be preventing it from working and breaking it. | 20:14 |
MovedMacBook2Dev | ok... the fedora gnome - nm ui - via gnome settings keeps spinning for a long time.. and then displays a notification via gnome UI that - Connection failed - Activation of n/w connection failed | 20:15 |
rwp | We can see the ps you posted earlier with the command line NM used to invoke dnsmasq and I don't see anything scary there. But if it isn't working then it is probably something else that is preventing it from working. | 20:16 |
MovedMacBook2Dev | previously the ip assignment to clients were almost instant / negligibel | 20:16 |
MovedMacBook2Dev | i see | 20:16 |
rwp | My next thing to check would be firewall rules. Because those block things by design. | 20:16 |
MovedMacBook2Dev | wow | 20:17 |
MovedMacBook2Dev | oh my god | 20:17 |
MovedMacBook2Dev | so true | 20:17 |
MovedMacBook2Dev | i installed firewall day before yesterday | 20:17 |
MovedMacBook2Dev | ufw | 20:17 |
rwp | There you go. | 20:17 |
MovedMacBook2Dev | and setup some rules | 20:17 |
MovedMacBook2Dev | to allow only ports like 80 and 443 | 20:17 |
MovedMacBook2Dev | this is it - - ufw status - https://dpaste.org/HveFO | 20:18 |
MovedMacBook2Dev | i honetly didnt think that will play its ways with sharing the ethernet connection as wifi ap hotspot | 20:19 |
rwp | That definitely will be blocking DHCP then. | 20:19 |
rwp | And it probably is interfering with NetworkManager setting up routes for the connection sharing. | 20:20 |
rwp | If it were me I would disable ufw and reboot and see if everything works. | 20:20 |
MovedMacBook2Dev | wow.. good call.. let me whitelist those necessary ports | 20:20 |
rwp | Firewalls are a good thing so I am NOT advocating for not using a firewall at all. Just isolating this problem now. And then carefully adding the firewall again with this in mind. | 20:21 |
MovedMacBook2Dev | yes .. disabling worked OOB - connected immediately | 20:21 |
rwp | Yay! \o/ | 20:21 |
MovedMacBook2Dev | \oz | 20:22 |
MovedMacBook2Dev | \o/ | 20:22 |
MovedMacBook2Dev | next steps is to figure out necessary ports.. fedora connected immediately | 20:22 |
MovedMacBook2Dev | lol | 20:22 |
MovedMacBook2Dev | so did my phone | 20:23 |
MovedMacBook2Dev | private dns wasnt even an issue | 20:23 |
rwp | You will want to allow ports 53 UDP/TCP, and 67, 68 for both UDP and TCP. | 20:24 |
MovedMacBook2Dev | reboot wasnt even neede | 20:24 |
rwp | On the private network side of thigns. | 20:24 |
rwp | I don't know how to do that with ufw. I am not using it. I am a Shorewall user. | 20:24 |
rwp | Everyone uses the new fangled stuff and I never use the new fangled stuff so I can never help people with the new fangled stuff. Sorry! | 20:25 |
rwp | DoT uses port 853 so if you have configured your system as a DoT server (you mentioned that earlier) then that would also be needed. | 20:26 |
rwp | And with that I am heading to do a work sprint on $DAYJOB for another bit. Good luck! :-) | 20:26 |
rwp | I usually configure iptables (Shorewall) to log as info anything that is blocked by the firewall on the private LAN side of things. Then I look at the logs and see what I am blocking. Sometimes that leads me to things I need to enable. Sometimes to clients on my LAN that are doing bad things that need to be reconfigured. | 20:28 |
rwp | Maybe ufw has such a feature too? I don't know. | 20:28 |
MovedMacBook2Dev | thanks so much rwp and gnarface - | 20:31 |
MovedMacBook2Dev | this helped me lot.. | 20:31 |
gnarface | np | 20:31 |
MovedMacBook2Dev | and also updated on me on daemon logs and syslogs as well | 20:31 |
MovedMacBook2Dev | so thanks for that as well | 20:31 |
MovedMacBook2Dev | thanks all again :) good day ! | 20:38 |
gnarface | heh, if you add that daemon.log line back to the rsyslog.conf, don't forget to also add it back to /etc/logrotate.d/rsyslog | 20:45 |
hagbard | Hi, can't get lightdm to run anymore. "Failed to use bus name org.freedesktop.DisplayManager" | 21:33 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!