[00:02] sarnold Plymoth should be as it was a functioning install for several months prior. How would I go about rebuilding initramfs? I currently have it booted from a USB stick while I'm typing this on my main system. [00:03] intok-work: chroot into the mounted system, bind mount in /proc /sys and /dev , then .. mkinitramfs? I think [00:14] sarnold assume I'm an idiot, this is all new stuff to me [00:14] any idea why ubuntu arm doesn't detect the storage on the raspberry pi 4? [00:14] intok-work: ah, sorry :) [00:15] the net install can't detect the memory card [00:16] and when I do a cchroot build it won't start. [00:16] chroot using debootstrap [00:16] intok-work: once you've got the root filesystem mounted on eg /mnt you would do mount --rbind /dev /mnt/dev ; mount --rbind /proc /mnt/proc ; mount --rbind /sys /mnt/sys ; chroot /mnt /bin/bash --login [00:16] intok-work: (I copy-pasted those from one of the blocks here https://github.com/zfsonlinux/zfs/wiki/Ubuntu-18.04-Root-on-ZFS#rescuing-using-a-live-cd ) [00:20] I did that but trying to boot the pi 4 with the install didn't work [00:22] sarnold so it's mounted [00:22] the boot loader is up but nothing happens after that. [00:22] cause it's not detecting the mocrosd card [00:23] intok-linux: alright, nice; and once you're chrooted, ps and top and the like seemed to work fine? [00:23] is there One True Way to manage CUDA (and nvidia driver) installation on ubuntu? i always end up with trouble... just now pytorch is complaining that my cuda driver is out of date but it was working fine before. i think that an ubuntu auto update nuked my cuda driver and/or nvidia graphics driver. [00:23] intok-linux: if they work fine then the mount --rbinds should be working... [00:25] intok-linux: I'm hoping/guessing that rebuilding the initramfs may help. re-installing grub may help [00:25] if you're on about the pi it doesn't use grub.. it uses uboot. [00:26] intok-linux: to update the initramfs, try update-initramfs -k all -u -- and I think update-grub will rebuild your grub config [00:26] raidensnake: intok-linux's got a laptop, not pi ;) [00:27] I'm poi nting out I can't get ubuntu arm to work. [00:27] cause it can't detect the storage. [00:27] raidensnake: my pi3b just worked, I didn't have to do anything funny.. I don't know how it works well enough to give any advice. if you could get the kernel booted, then we'd be some place where I can have ideas :) [00:30] @sarnold I'm using a pi 4. [00:31] I'm just saying ubuntu won't boot on a pi 4 [00:34] Hi I'm having trouble getting audio on a creative ca0132 chipset does anyone have something that works? I've tried several things and nothing has helped [00:35] when I reload the alsa drivers ca0132 is listed as one that's reloaded I just don't know why it won't play sound [00:46] i've tried https://askubuntu.com/a/472769/971074 to send notification via cronjob but no luck what did i miss "*/2 * * * * export DISPLAY=:0 && /usr/bin/notify-send -t 12000 BO" [00:47] why not use a gui for chron [00:48] gui apps for cron Sveta [00:48] sorry I rebooted, did anyone respond to my sound issue? [00:48] freeitt: sorry, no [00:49] dang, not sure why I'm having issues after reinstalling. I had no issue with a normal installation before this [00:49] thanks though [00:52] chieta, yes, i mean gui apps for chron [00:52] chieta: did cron send any emails that might include errors? [00:55] here i used cli Sveta [00:55] no, it didn't sarnold [00:56] the command works but it doesn't on cron [00:56] what about...\ [00:56] DISPLAY=:0 /usr/bin/notify-send -t 12000 BO [00:56] as one command, not two [00:56] heh, good idea [00:59] checking... [01:03] im on 19.04 and it doesn't work... but i see the job executed and logged on syslog [01:05] chieta: hmm. maybe add to the end... > /tmp/output 2> /tmp/error -- see if you can find out more that way? [01:08] what's the "best" sftp server for linux/ubuntu? [01:08] sshd [01:08] better than sftp? [01:09] openssh server provides sftp [01:09] i mean vsftpd [01:09] and sftp does not work without ssh [01:09] vsftpd is not sftp, it is another security/auth method [01:09] gotcha, so is openssh = sshd? [01:10] sftp stands for 'SECURE ftp' and involves authing over the ssh protocol. ftp does not use that and uses its own username+password pair instead [01:10] yes [01:10] ok so basically to setup an sftp server i just create an ubuntu install with openssh-server and configure it [01:10] sftp stands for SSH File Transfer Protocol [01:10] becool: yep [01:11] thanks [01:11] dax, thanks for the correction [01:11] (confusingly, sftp is unrelated to ftp or to ftp-over-TLS aka ftps. the latter two are what vsftpd would get you) [01:13] sarnold no output and error listed [01:14] did the command work on your env, sarnold? [01:14] Sveta [01:15] chieta: good question, let me try [01:17] chieta: yeah, I added this to my user crontab with crontab -e and it worked flawlessly: * * * * * DISPLAY=:0 /usr/bin/notify-send hi [01:18] noted, sarnold i will try without -t switch === zbenjamin_ is now known as zbenjamin [01:29] Ok, new question, how do I force the ca0132 driver onto the card instead of using the snd_hda_intel I want to use snd-hda-codec-ca0132 [01:35] freeitt: hmm I think I've seen options like that in a gnome sound control panel once [01:35] freeitt: *maybe* pavuctrl? or how is that spelled.. [01:37] I haven't seen anything like that in either of those programs [01:40] ;( [02:20] Hello, how can I tell if I need another wireless driver for my WiFi card? It keeps randomly dropping [02:21] How to pass kernel parameters to a process during run time [02:22] Using ubuntu desktop 19.04 [02:22] dsag: what's a "kernel parameter" that you would pass to a process? [02:23] Nothing specific as of now. But Anything parameters [02:24] @setharnold you could maybe give an example? [02:24] dsag: well, that's why I asked you :) I don't know what you're trying to do [02:25] Lets say if i want to allocate more memory to a process [02:25] Can we pass some kernel parameters? [02:26] dsag: processes allocate memory as they need themselves; do you mean modifying rlimits? cgroups? or something else? [02:26] Lets say if i want to modify [02:29] I am using an Intel Wireless 8260 according to lshw -c network [02:29] What drivers should I be using for that [02:32] xrandr_: you could check the output of modinfo /lib/modules/`uname -r`/kernel/drivers/net/wireless/intel/iwlwifi/iwlwifi.ko to make sure that your hardware matches one of the alias: lines from that module [02:36] I guess not [02:36] Wired it is then, hopefully [02:37] brb === [b0b] is now known as b0b [02:58] Anyone around who could help with an fstab parse error? [02:59] Casper26, better to describe your problem and you may get someone helping. [03:00] get a parse error when adding s23dc01:/DOCS /media/s23dc01/DOCS nfs auto nofail,uid=1000,gid=1000,noatime 0 0 to fstab? [03:02] Casper26, try it without space between auto & nofail (they are the same paramater & should be seperated by comma) [03:05] @guiverc that worked thanks alot! [03:26] How well does Ubuntu 18.04 LTS run on a 2009 core 2 duo with 2 gigs of RAM? [03:26] Is it even worth borhtering to try and install it? [03:27] I don't mind switching DE's if necessary. [03:27] Just trying to do some web browsing and maybe some C programming on it, nothing huge. [03:28] AlexMax, It should run okay. But with 2G of ram the limiter across all systems will be web browsers. Firefox and Chromium are pigs. [03:28] Aside from those everything else you mentioned wanting to do should be fine. [03:35] rwp: Well, 8 gigs of RAM is $37 off newegg [03:35] but for $37 I could probably get a better laptop lol [03:36] True. I paid a total of $89 for the x220 I am typing on now. With 8GB of ram. [03:36] ...tell me more :P [03:36] Mine is a T500 [03:36] What's there to tell? eBay. Shop for what you know. [03:37] But in the meantime. I say go for it with the installation. It might be just enough for you. Why not? [04:06] j [04:09] Welp, it's installing, giving lubuntu a shot [04:28] Fun keyserver-related question that probably belongs here. [04:28] My GPG key was signed by some spammers, and I would like to remove the spam signatures. [04:28] I currently hold my GPG private key, so I hope this is still possible. [04:29] I haven't been able to find working documentation on this yet. [04:29] The GPG key in question is the one on my LP profile (~tsimonq2). [04:30] how can a spammer sign your gpg key without having your privat key? [04:30] ¯\_(ツ)_/¯ [04:30] Well, they can sign my key. [04:30] I didn't sign theirs. [04:31] They pushed my key directly to the server. [04:31] When I push my own key directly to the server, it doesn't overwrite the spam entries, which is "wat" in and of itself. [04:32] tsimonq2 Sveta could this be related? https://www.vice.com/en_us/article/8xzj45/someone-is-spamming-and-breaking-a-core-component-of-pgps-ecosystem [04:32] lotuspsychje: Nope. My key was spammed a while ago, that's unrelated. [04:32] kk [04:32] I'm just thinking of this now. :P [04:34] https://keyserver.ubuntu.com/pks/lookup?op=vindex&fingerprint=on&search=0xE27F2CF8458C2FA4 [04:34] e.g. "Matthew Knight " [04:38] tsimonq2: sure its not this? just brainstorming with you https://access.redhat.com/articles/4264021 [04:38] you can't remove signatures (or anything else) from SKS keyservers [04:39] Aww man. [04:39] malicious signed, that sounds like what you have right? [04:39] Correct, but like I said, these signatures were from 2018. [04:39] dax: Even if I have the private key? [04:39] That's disappointing. [04:39] yes, even if you have the private key [04:40] it's inherent in the design of SKS, it's not a new thing (people bothering to do it tens of thousands of times is a new thing) [04:40] I'm curious what the rationale behind that is. [04:40] I can see not removing keys if you don't have the private key or revocation cert. [04:41] That's fine and dandy. [04:41] https://gist.github.com/rjhansen/67ab921ffb4084c865b3618d6955275f#keyserver-design-goals [04:43] Well, thanks anyway. [04:43] tsimonq2: nice git here too: https://gist.github.com/rjhansen/67ab921ffb4084c865b3618d6955275f [04:43] seems like what dax said is mentioned under concequences [04:43] Yeah. [05:12] REEEEEEEEEEEEEEEEE [05:12] ok [05:26] Greetings everyone. The question I come with tonight it more theoretical than anything else, and may well be beyond the scope of a support channel. So, with that said, lets say I have a file that I want generated. And lets say that I want to be notified of that by an application window. Not a standard notification, but a true application window, something that would pop up in my panel/task manager/ whatever you want to [05:26] call it. Would this be possible? [05:27] Of course, I'd be checking for the file with a shell script [05:28] I suppose similar to how a message box gets generated with a windows VBscript, but actualy with a window that was on my panel [05:28] So i need help how do i point my domain name to apache on ubuntu [05:30] adol-christin: Provided the service is running and ports are open, it should be as easy as pointing the, I think it's a c-name record(?) to your public IP [05:30] im using google domains [05:31] for my domain [05:31] i had to format my dedi because i played around with the host settings and it kept crashing over and over [05:31] ye ports open [05:31] can connect to my website via ip but not domain can u maybe help a poor young man? [05:32] adol-christin: did you reformat with ubuntu server LTS now? [05:32] ye about 20 minutes ago [05:32] just installed apache [05:32] adol-christin: ok, join #ubuntu-server please [05:40] do why can't i have helps here === imsurit_ofc is now known as imsurit [05:43] adol-christin: the ubuntu support channels are divided for a reason, they help specific topics for specific issues/versions/topics [05:44] but its dead :( [05:44] its deader than a dodo [05:44] adol-christin: its not dead, its friday morning, a little patience please [05:44] Ubuntu Server is on-topic in both #ubuntu and #ubuntu-server, just like official flavors are on-topic in both their flavor channels and #ubuntu. [05:44] its friday arvo [05:45] im not american === morphis6 is now known as morphis [06:15] hello. I am running bionic. if I connect an external USB Hard drive, how can I find out if it is USB 3.0 or 3.1? Also, how can I find out if my computer supports USB 3.1? thank you [06:19] Mr_Cyclops: not entirely sure, but dmesg can probably tell you. that's how to tell whether USB 2.x or 3.x is supported. [06:20] Mr_Cyclops: as for the drive, it's hard to say. most don't indicate what level of USB they're currently using, so either checking dmesg or the IO rate is probably the only option. it also depends on what port they're connected to at times. [06:20] ryuo, thanks. though I just figured it out. Found a nice link [06:20] not sure if its allowed to post links here in the forum? lotuspsychje ? [06:21] i do have an enclosure that has an LCD that shows what mode it's currently using, but that's a rarity. [06:22] get the Bus ID and Device ID from lsusb output. then do this [06:22] lsusb -D /dev/bus/usb// | grep bcdUSB [06:24] though i honestly can't see 3.0 -> 3.1 making a huge difference for external drives, unless they're nvme. [06:24] if they're plain sata, then they'll be bottlenecked more by SATA.. [06:26] true, 3.1 is way faster [06:26] though nvme is bizzarely (is that a word?) fast! [06:26] (links are fine as long as they're on-topic, i.e. related to support) [06:26] dax, thank you :) [06:27] so ryuo .. the link is https://askubuntu.com/questions/604158/how-do-i-tell-if-a-usb-thumb-drive-is-usb-3-0 [06:28] has anyone tried this new fcitx5 package? [07:06] Beta channel: If you are interested in seeing what's next, with minimal risk, Beta channel is the place to be. It's updated every week roughly, with major updates coming every six weeks, more than a month before the Stable channel will get them. [07:06] updated once or twice weekly, and it shows what we're working on right now. There's no lag between major versions, whatever code we've got, you will get. While this build does get tested, it is still subject to bugs, as we want people to see what's new as soon as possible. [07:06] uild has not been tested or used, it's released as soon as it's built. [07:06] ^ Which one is best? [07:07] I'm thinking Dev is a safety between Beta and Canary [07:18] I am experiencing strange problem on my laptop running 18.04 . Just a hover on any icon or link click event occurs and open the things automatically without any click. [07:19] I have reinstall ubuntu but experiencing same problem. What could be the cause? [07:19] mouse issue [07:19] or maybe a stuck key [07:19] try disabling onboard inputs [07:20] I disabled trackpad but no luck [07:25] I removed click buttons and enter key from laptop but facing same issue [07:26] Can humidity cause such issues? [07:31] are you experiencing the problem with any other OS? [07:31] If not, it's a hardware issue. [07:32] fosserjosh: is the same happening in a liveusb? system up to date? [07:34] System is up to date. Even for liveusb same issue [07:36] fosserjosh, hardware. deep your track if possible. [07:36] or switch to external mouse [07:36] and test an external keyboard === mnemonic is now known as Guest49852 === EriC^ is now known as EriC^^ [07:48] @cfhowlett can i disable laptop keyboard? How? [07:48] wait 1 [07:48] I used external mouse but with that also same issue [07:49] Blue tooth keyboard is what I use. [07:50] But with Bluetooth keyboard laptop keyboard still works [07:50] confirm: no problem with the bluetooth keyboard? [07:51] I am not using Bluetooth keyboard [07:51] This looks like some laptop internal issue or laptop keyboard issue [07:52] right. connect a bluetooth KB to test. [07:53] Now Bluetooth keyboard connected still same issue [07:54] internal laptop issue. keep asking in channel to problem solve this. [07:55] Can humidity play role in this? Currently climate here is humid [07:55] Due to rains [07:56] .... yeeeeeeeeeeees?? [07:56] interesting problem! [07:56] but, of course, you should always protect against moisture. [07:57] and if it were moisture I suspect you would be seeing multiple issues not just this one [07:58] I can see line on screen too [07:58] Sometimes [07:59] new laptop [07:59] ?? [07:59] Nope its been 7-8 years [07:59] ah. OLD laptop. actually, that might be the issue. [07:59] Experiencing 1st time such issue [08:00] I had erratic behavior on my old compaq lappy. turned out the hinge spring came loose and was shorting out on the mobo. [08:01] one more test: alternate OS. try xubuntu or lubuntu. [08:01] Even my laptop also having issue with one of the hinge [08:01] Will check that [08:01] might be time to upgrade the hardware, amigo. [08:02] True that [08:04] Is there anyway i can see log about this ? [08:04] Like click event [08:05] dmesg or the log viewers but I have no idea exactly what to look for nor how to interpret it. ask in the channel [08:06] dmesg | more for one screen at a time === The_Milkman1 is now known as The_Milkman [08:38] My apps suddenly don't connect to the internet except my browsers. [08:39] Slack, Hexchat, Discord, Telegram, and even the `apt update` command don't connect to the internet. [08:39] What could be responsible? [08:43] newhorn: describe "don't connect", what error exactly do you see in apt for example [08:49] blackflow, can I PM the error? I can't even connect to a pastebin site. [08:49] Only Google searches and some sites work. [08:50] newhorn: I'm guessing it's DNS. You can temporarily add "nameserver 8.8.8.8" at the top of /etc/resolv.conf (without the quotes, see the other nameserver entry in teh file), and see if that fixed your situation. [08:51] nameserver 127.0.0.53options edns0 [08:51] nameserver 127.0.0.53 [08:51] options edns0 [08:51] Those are the current entries. [08:52] right. so add 8.8.8.8 at the top [08:53] blackflow, Do I add nameserver before that? [08:53] "nameserver" [08:53] newhorn: re-read what I wrote please. [08:54] blackflow, Oh your earlier message answers that. My bad. [08:57] newhorn: aaaaand did that fix your immediate issue with DNS? [08:57] blackflow, Do I need to restart something after this change? [08:58] newhorn: nope [08:59] Okay. Issues persist. [09:01] newhorn: okay. can you ping -c1 termbin.com ? do you get "Name or service not known"? [09:03] blackflow: 1 packets transmitted, 0 received, 100% packet loss, time 0ms [09:03] newhorn: try again with -c 10 [09:04] 10 packets transmitted, 0 received, 100% packet loss, time 9202ms [09:04] newhorn, traceroute 8.8.8.8 [09:04] newhorn: and via the browser? Can you access termbin.com ? you shouldn't but.... you mentioned browsers "worked" [09:05] blackflow: Browsers work, so I am able to use webchat.freenode right now, and Google searches work, but not most sites. [09:05] Mathisen: why 8.8.8.8 the test should be consistent, we just established termbin.com is unaccessible [09:05] Mathisen: traceroute command not found [09:05] blackflow, to see where it stops ?? [09:05] newhorn: and can you please answer my question? [09:06] blackflow: can't access. [09:06] Gmail also works. [09:06] Mathisen: they just said some sites works, some don't, so tests should be consistent. if there's 100% packet loss to termbin.com, then that should be tracrt'd because what if 8.8.8.8 works? that result doesn't tell anything [09:07] newhorn: paste.ubuntu.com? dpaste.de? in the browser [09:08] Both don't work. [09:09] newhorn: tried to restart your upstream router? also can you post here the output of `ip route` ? It _should_ be three lines. if it's more, then please paste just the one starting with "default via" [09:09] default via 192.168.8.1 dev wlp2s0 proto dhcp metric 600 [09:09] 169.254.0.0/16 dev wlp2s0 scope link metric 1000 [09:09] 192.168.8.0/24 dev wlp2s0 proto kernel scope link src 192.168.8.102 metric 600 [09:10] newhorn: can you ping 192.168.8.1 ? [09:10] What does this mean? "tried to restart your upstream router?" [09:11] blackflow: Ping works. [09:11] install net-tools? [09:11] newhorn: the machine at 192.168.8.1, that's your router in the LAN I assume? Did you try restarting that? But also, I think you have a problem with that wifi, the route doesn't look natural. you shouldn't have link-local set up [09:12] I use a wifi device (modem). [09:12] Perhaps, I should restart that? [09:13] newhorn: yes. stop your wifi connection, restart the "wifi device (modem)", and start your wifi connection again. That will disconnect you from here I suppose. [09:13] blackflow: Yes. Brb. [09:17] blackflow, Works! [09:17] Restarting did the magic. Thanks! [09:17] Should I leave the "etc/resolve.conf" content as is? [09:18] newhorn68: you can remove that "nameserver 8.8.8.8" line and leave it as it was. [09:18] keep in mind he had you change your dns to google one newhorn68. you are using that one now as default [09:19] Okay. I'll change it back to that. Thanks guys! [09:19] Oh. The system already changed it back to that. [09:20] probably, when you restarted the wifi, as resolv.conf is by default a dynamic file under /run/, under jurisdiction of NetworkManager [09:22] Makes sense. [09:26] Hi! Is there a way to remove a hardlink from a file, but preserve the file? This means that the file must be copied. "rm" deletes the file [09:28] regdude: are you asking how to convert a file that's a hardlink, into a separte copy of the file it was a hardlink to? [09:30] regdude: any file is a hardlink really, when all 'hardlinks' are removed, the inode becomes free and the space can be used [09:31] if you use ctrl-z [09:31] How do you get up the current job in tarminal again? [09:32] Whiskey-: type 'fg' [09:32] and how do you select what process it should be that is put back in fg? [09:33] fg PID seems to not work [09:33] EriC^: that's a bit misleading. You're referring to inode references, yes, but there's no copy on write going on at least not with ext4 afaik. if you copy -l one file to another and then modify the original? what happens to your hardlinked one? [09:33] Whiskey-: type 'jobs' , then type "fg %" [09:35] blackflow: yes, I would need to create a copy, but not sure how I could do that without creating a temporary file, removing the link, then renaming. Isn't there a single command for that? [09:36] regdude: not that I know of. you need a separate inode, so a proper copy (with no -l) is in order. [09:36] blackflow: i'm not sure what you mean, if you make a hardlink of a file, then modify the hardlink? the original will change since it's just a hardlink [09:36] EriC^: I mean there's no copy-on-write mechanics here. say you have a file A. You write "hello" in it. then you cp -l A B. then you write "world" in A. What's teh contents of B? [09:37] blackflow: it should be 'hello' [09:37] EriC^: no it's not :) [09:37] there's no copy-on-write going on. [09:37] blackflow: that sucks, well thanks for confirming. Will try to figure out if sed is not capable of doing so [09:38] blackflow: um, you're mistaken :) [09:38] EriC^: well try it, don't believe me. echo "Hello" > A ; cp -l A B ; echo " world!" >> A ; cat B [09:39] blackflow: https://paste.ubuntu.com/p/dyz5HVVdYq/ [09:39] EriC^: yes you just confirmed what I said [09:39] lol what [09:39] you echo in bla and your blalink has the same contents [09:39] same as bla [09:40] exactly [09:40] which is what I said. with hardlinks, you hardlink reflects the original inode. if you modify the contents of orig file, the hardlinked one reflects that because it's not a copy on write [09:40] i misread your last statement, about the hello and world (i didnt see the world part), but i did mention earlier that A would change.. [09:40] no you said "it should be hello" while it should be "hello world". [09:40] blackflow: yes i know, that's exactly what i said [09:40] but anyway, let's not argue [09:40] read above [09:40] lol [09:41] the point is, regdude wants to "detach" the hardlink. they must make a proper copy. [09:41] "i'm not sure what you mean, if you make a hardlink of a file, then modify the hardlink? the original will change since it's just a hardlink" [09:41] i dunno what the problem is [09:41] typical #u stuff help a person out get into a fight with regs [09:41] EriC^: put me on /ignore and gtfo I'm sick of your BS [09:42] calm down buddy [09:42] you're misreading what im saying and talking about stuff from the air and arguing and then you're the one who's sick of me? ok [09:42] lol [09:43] the only argument here is started by you, "misunderstanding" what I wrote from the beginning. ther is no copy on write. If you have no clue what that means, that's your problem. regdude wants a proper copy. that's all. [09:43] blackflow: i didnt even say anything except explain to the person what a hardlink actually is [09:44] EriC^: just put me on /ignore and case closed. [09:44] I probably should give a backstory. I used a file backup for a service, but I made a mistake, hardlinks were ignored when backup was restored. I used rdfind to find duplicates and fix the missing hardlinks, but that created a problem with few files that should not be linked together although they are the same. If one link end changes, then the other link end changes, I need to avoid that for a few files. The solution is to copy th [09:44] but doing that is a bit tricky (not hard, just frustraiting) [09:45] blackflow: maybe you should put me on /ignore [09:45] no, I want to correct BS advice when I see it. your statement "they're all hardlinks" is misleading and wrong in this case. [09:46] blackflow: they're all hardlinks? what are you actually saying, listen im trying to tell him what a hardlink actually means that's all [09:46] regdude: yeah I don't know of another way as you literally need to create a separate inode entry, which is done by a regular copy (I'd use cp -a) [09:47] EriC^: "11:30 < EriC^> regdude: any file is a hardlink really," -- no it's not. [09:47] blackflow: what? [09:47] lol [09:47] blackflow: "any file is a hardlink really, when all 'hardlinks' are removed, the inode becomes free and the space can be used" [09:48] what's the problem with that sentence? if you have 1 file that exists, and you stat it, you'll see it has 1 link aka hardlink, which is the one you're stat'ing [09:48] that statement is misleading, conflating "inode references" with "hardlinks" [09:49] what? [09:49] you have problem reading? [09:49] !discuss [09:49] Want to talk about Ubuntu, but don't have a support question? /join #ubuntu-discuss for non-support Ubuntu discussion, or try #ubuntu-offtopic for general chat. Thanks! [09:49] dont do that here guys [09:50] there's nothing wrong with my sentence, i said the file is a hardlink (even just 1) and you can make other hardlinks, when all hardlinks are removed, the inode is freed and you get the space back [09:50] if you're reading stuff into it and whatever, it's just you blackflow [09:50] im done [09:58] regdude: if they are text files, GNU sed's -i will create a new file and rename it over the original, effectively severing the link. May corrupt non-textual files though [10:17] geirha: sed -i '' file.txt <-- does this seem right? [10:18] regdude: that should do it, yes. Might add a trailing newline if the original didn't have one [10:19] regdude: I wouldn't use sed. it's not that hard to script cp -a to work through /tmp/ [10:21] ok, that should be the safest way since there might be a few binary files. Thanks for all the help everyone! [10:23] I'd copy to a temporary file in the same dir, so the following move becomes atomic [10:23] good advice. cp followed by mv [10:30] find . -type f -links +1 -exec sh -c 'cp -i "$1" "$1.tmp" && mv "$1" "$1.tmp"' sh {} \; [10:30] hi [10:30] just like backlight file are there any other files that access monitor? [10:30] I'd like to know about more such /sys/ or other files if any? how would I get a list of them? [10:36] hm.. [10:36] i can't kill postgres process. [10:36] and it's loading 100% cpu [10:37] and excuting bash64 [10:37] kill -9 [postgres_pid] .. it immeidatley executed again. [10:43] anybody is there? [10:44] !patience | monkeyisl [10:44] monkeyisl: Don't feel ignored and repeat your question quickly; if nobody knows your answer, nobody will answer you. While you wait, try searching https://help.ubuntu.com or https://ubuntuforums.org or https://askubuntu.com/ [10:45] monkeyisl: details might also be useful, ubuntu version etc [10:48] lotuspsychje : 16.04.5 LTS [10:50] monkeyisl: desktop or server? [10:50] desktop [10:51] lotuspsychje : desktop [10:51] No LSB modules are available. [10:51] Distributor ID : Ubuntu [10:52] !paste | monkeyisl [10:52] monkeyisl: For posting multi-line texts into the channel, please use https://paste.ubuntu.com | To post !screenshots use https://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [10:52] description: Ubuntu 16.04.5 LTS [10:52] please update your system, 16.04.6 is current [10:53] https://imgur.com/a/LxMk8HF [10:53] how do i update? [10:54] oh cchking [10:55] new release '18.04.2 TLS' available. [10:55] 'do-release-upgrade' to upgrade to it [10:56] should i go? [10:56] monkeyisl: the users choice, do you want 18.04 LTS upgrade, or just update xenial to .6? [10:57] i'll try 18.04.2, but unsure if it will fix the problem though. [10:57] updaeting [10:58] monkeyisl: its adviced to make backup of your data before LTS upgrading [10:59] monkeyisl: i meant update, as in sudo apt update && sudo apt full-upgrade [11:00] this would install the missing security and bug fixes on your system [11:00] but you can also continue doing what you started now, doing a release upgrade to ubuntu 18.04. but even then, you need to ensure to keep this new version updated. [11:01] (but i agree that you should always have complete and proven restorable backups at any time, and especially before an upgrade) [11:14] do you know what's linux hooks to change monitor from a low level backend ? [11:26] hortiel: you'll need to provide more context [11:26] Hey folks [11:26] tomreyn: are there any low level access similar to /dev/sda [11:26] that's if this is an ubuntu support rather than a generic linux programming question [11:27] so that I can get some low level control similar to device files!! [11:28] well, monitors are very different devices than storages :) [11:29] as a result, interfaces differ a lot, too. maybe ask in #linux - probably a better place than this ubuntu support channel. [11:29] * ##linux === jstein_ is now known as jstein [12:44] Hi, I can't move anything to Trash Can on auto mounted ext4 partition, any ideas? [12:46] hi guys when i try to install nmap a program i get following output from apt-get https://paste.ubuntu.com/p/9KsDscGWjD/ [12:47] funabashi: 17.10 is end of life [12:47] funabashi: install a supported ubuntu version from the topic please [12:47] please i dont want a new verion for this pc [12:48] funabashi: the problem is that ubuntu wont run mirrors for that version of ubuntu anymore [12:48] so your apt is going to be pretty broken [12:48] nobody? [12:48] Gosset, check the permissions (and mount options too) [12:48] mount options? [12:48] yeah [12:49] what file [12:49] funabashi: in a pinch, you could edit /etc/apt/sources.list and change the domains to http://old-releases.ubuntu.com/ubuntu/ - but you should consider upgrading to 18.04 which is LTS https://help.ubuntu.com/community/EOLUpgrades [12:50] Gosset, you can use gnome-disks ( if automounted) [12:51] the options are: nosuid,nodev,nofail,x-gvfs-show,x-gvfs-name=Data [12:52] Gosset, you can add 'user'; iirc, just opening nautilus creates a trash-like folder on the mounted partition, so you can ls -al it to check the permissions [12:53] I'm on Ubuntu Mate [12:53] rory: its seems to be down how can i change it to ipv4 or instead of ipv6? [12:54] anyway, in Caja it's the same? [12:56] I'm googling it but most of the topics are for NTFS partitions [12:56] mine is ext4 === Flexman is now known as Guest77505 [12:59] what if I replace nosuid by uid=1000 ? === zbenjamin_ is now known as zbenjamin [13:05] funabashi: why would you not want to run a supported release? [13:06] and why would you want to use software with exploitable security vulnerabilities? [13:06] rory: now this error https://paste.ubuntu.com/p/CNZNKcZz2W/ [13:06] tomreyn: its my job [13:06] ... "debian jessie"? [13:06] the PPAs you used also don't provide packages for unsupported ubuntu releases. [13:07] those errors shouldn't prevent you installing nmap [13:07] funabashi: your job is to run vulnerable software? [13:07] you could remove the PPA and the debian jessie repo from sources.list [13:07] tomreyn: pentester [13:07] i see [13:07] ioria ? [13:08] funabashi: so you're building a target to run tests against? [13:08] kind of yes [13:08] you said I could add "user" to options? [13:08] in nosuid,nodev,nofail,x-gvfs-show,x-gvfs-name=Dades [13:08] ?? [13:09] but then you wouldn't need nmap on it, you'd want that to run on a fully patched system, i guess [13:09] funabashi: that's one specific target there... that it requires all sorts of outdated PPAs, too - why not pentest a regular (outdated) system? [13:09] thanks anyway [13:09] Gosset: what is the original issue? that it can't create the trash directory in that filesystem? [13:09] please focus on my issue instead of nmap, apt-get doesnt work now [13:09] yes [13:10] funabashi: EOL releases ar enot supported here, you're on your own. [13:10] there's no trash can [13:10] cat /etc/debian_version [13:10] stretch/sid [13:10] on my auto mounted partition [13:10] which is ext4 btw [13:10] not ntfs [13:11] Gosset: you'll have to create it manually, it seems (see https://specifications.freedesktop.org/trash-spec/trashspec-1.0.html) [13:12] manually? [13:12] :P [13:12] Gosset: the trash directory is usually created inside $XDG_DATA_HOME, or at the root directory of a filesystem if the user's home is not on that filesystem [13:12] ok [13:12] thanks [13:13] it's strange that if it was NTFS, I could solve this matter by adding just a word in the partition options [13:13] Gosset: that's because linux does not understand NTFS file ownership/permissions, so it treats all the stuff in an NTFS as "root-owned" (or "specific-user-owned" with the appropriate flags) [13:14] for ext4 (or any other POSIX-compliant FS), this is not necessary, as all the permissions are understood correctly [13:14] however, this leads to the situation where you can't reall "mount and ext4 filesystem as this-or-that user" [13:14] s/reall/really/ [13:15] I understand [13:15] if you want to give your user write permissions to the top-level of that filesystem, you'll have to chown it - but that might not necessarily be a good idea [13:15] (depending on how you intend to use that drive otherwise) [13:16] funabashi: your original issue was that you wanted to install nmap - can you install nmap now? [13:17] ayekat: volunteers have already adviced not to support EOL versions, please respect that [13:17] lotuspsychje: well, they kept on asking [13:18] nothing you can do anyway, 17.10 is deader than a door knob [13:19] It's dead, Jim. [13:20] ah, they've moved on to ask in #debian now anyway ^^ [13:21] aye [13:29] ayekat: nope [13:29] funabashi, if the system was not actually installed as jessie, the best option is to use repos from the archive of the actual distro. That would mean deb lines, not just deb-src lines. [13:30] lotuspsychje, does anyone here advise on helping a user move away from EOL to still-supported? [13:30] everyone did, they chose to ignore them [13:31] ¯\_(ツ)_/¯ [13:31] jelly: scroll up [13:31] jelly: EOL versions are not supported here [13:31] lotuspsychje, did you even read what I said? [13:31] hardly. :) [13:31] they hardly do [13:32] we have !eolupgrade for this purpose [13:32] it needs an overhaul, though [13:33] jelly: would you take the risk advising users to eolupgrade knowing its a security risk, their system could already be compromized [13:33] tomreyn: funny [13:33] funabashi: it still works well enough. but then you don't want to upgrade, do you? [13:34] !eolupgrade [13:34] End-Of-Life is when security updates and support for an Ubuntu release stop, see https://help.ubuntu.com/community/EOL for more info. Looking to upgrade from an EOL release? See https://help.ubuntu.com/community/EOLUpgrades [13:35] lotuspsychje, absolutely, I do that every day albeit for a different distro [13:36] jelly: tell me all about it in #ubuntu-discuss [13:36] i want to be able to download and install programs not update the system [13:37] funabashi: you can't. 17.10 is dead, the repos nonfunctional. Savvy? [13:37] blackflow: i did install a program yesterday [13:37] so yes it works [13:37] worked* [13:37] Hi. I have a problem installing RabbitMQ. I have followed the steps described here: https://www.rabbitmq.com/install-debian.html but apt update fails with the following error: https://paste.ubuntu.com/p/rBmBHKfTGc/ [13:38] funabashi: it may have worked yesterday - 17.10 has been unsupported for a while now - there is no "make it work without upgrading" solution [13:38] funabashi: from a gods know what PPA/third-party-repo? [13:39] the main 17.10 repos are no longer available, and as you've demonstrated through pastes, you're running a franken-meld of several frankenbuntudebians. [13:39] stretch/sid with PPAs and whatelsewasthere [13:39] guys i know that you know how to fix this , but guess you like to argue [13:39] Apparently putting 21.x should work but it doesn't. Was I supposed to specify a version for erlang? [13:40] blackflow: to be fair, ubuntu is always going to show {debianrelease}/sid, because ubuntu appears to base their releases off sid [13:40] funabashi, dude. NOT SUPPORTED. as in ... it's not supported. [13:40] funabashi: nobody wants to keep the solution away from you - it simply doesn't work anymore, there is no magick [13:40] tomreyn: why do you care about my life? [13:41] funabashi: that's not a topic for this channel === msalvatore_ is now known as msalvatore [13:42] ayekat: I know but that sources.list is a total mess :) [13:43] they haven't shown it to us, but... yeah, guessing from that apt output O_o [13:43] ok i give up for now. i use git instead of apt [13:44] that's even... wait..... how do.... [13:44] * cfhowlett not even going to entertain the concept [13:44] * ayekat ducks away as blackflow flips the table across the room [13:49] leni1: looks like their instructions are not / no longer correct. get help from the rabbitmq project on this, or use ubuntu's packages (which are the only ones we can support here) [13:49] https://www.rabbitmq.com/#support [14:01] Ok i figured it out and fixed it first google domains is putrid only works with their partners [14:02] wont work with your dedi [14:02] namecheap worked instantly [14:12] What's the equivalent of libc-dev for Ubuntu? [14:12] leni1, libc6-dev [14:12] libc-dev-bin ? [14:12] jelly: thanks [14:13] So leni1 I use apt libc "tab" to get a list of packages then I look and choose the correct one. [14:15] well more specifically apt install libc "tab" yes I want to look at 1800 packages [14:15] mTeK: erm I wanted to know the correct one and as you can see here: https://packages.ubuntu.com/bionic/libc6-dev unless you knew the right one, it will take quite a bit of hopeful guessing :) [14:19] Without gasoline and a match how would you diagnose a port flap using DAC's in bond to mlagged switches. It starts happening after 8 days after a reboot... I've changed firmware on cards, installed driver as dkms as manufacture says, changed kernels, changed DAC cable manufactures. [14:20] I'm thinking next step is to create more mlagged interfaces and move the servers into the new port channels to see if it's the first 4 ports of this switch. [14:20] If that doesn't solve it should I change OS or nic cards next? [14:23] mTeK: you might have more luck over at #ubuntu-server [14:23] It's randomly after 8 days happening on 4 servers. These 4 servers are some older supermicro's. I wouldn't think it would be the motherboard or pcie slot issue. The flap is only on one of the DAC's, the other interface seem stable but it's on a different switch. [14:24] thsnr: didn't know about that room [14:27] * cfhowlett still worried about diagnosis via gasoline and a match === igghibu_ is now known as igghibu === paul__ is now known as octav1a [16:18] hello. i'm looking for a 15" laptop that works very well with ubuntu and has no numpad. suggestions, please :) [16:20] zutat, we are not the yellow pages, just ubuntu technical support [16:20] there is a certified list, sure you can find one without numpad [16:45] both universe, multiverse and restricted are allready enable [16:50] ubuntu-mate-18.04.2-beta1-desktop-armhf+raspi-ext4.img [16:50] ayekat ubuntu-mate-18.04.2-beta1-desktop-armhf+raspi-ext4.img [16:51] oh, ubuntu-mate - lemme check [16:51] ayekat sorry [16:52] can i ask ubuntu-drivers to download all drivers, so that they can be loaded while offline? [16:55] bumblefuzz: just to check, though (while this awfully slow torrent downloads the image), does it work with the regular ubuntu version for you? [16:56] linux-firmware are installed om both installations as well [16:56] after all, I expect switching to the MATE interface to be just a matter of installing a couple of packages... [16:57] ayekat I don't know [16:57] ayekat I haven't tried [16:58] ¯\(°_o)/¯ [16:58] It is an RTL8168evl [16:59] ayekat I wouldn't expect the underlying boot files to change much though... [16:59] I always forget what valid email addresses are... I think it's /^[^@]+@[\w.-]+/ ... right? [17:00] bumblefuzz: well, the fact is, on the regular ubuntu image I previously downloaded, the boot files were alright - so either something is weird on your side, or the files *are* indeed different in the MATE version [17:00] ash_worksi: This isn't a developer channel, please find a more appropriate channel for your question. [17:00] ash_worksi: the fully correct regex would take a *lot* more - but yeah, probably not the right channel here [17:01] ash_worksi: And the answer to your question is, there is no magic way to validate an e-mail address. The only bullet proof wya to validate e-mail is to send a test e-mail to th account and ask the user to click a validation link. [17:01] ash_worksi: entire sites dedicated to the subject. https://emailregex.com/ and then of course RFC5322 [17:01] and even they admit it doesn't work 100% [17:05] ayekat I'll just wait to see if your mate download checks out... otherwise, you might have more insight on the file differences than me [17:06] I tried installing r8168-dkms ... seems to be a resonable thing to try, it will take me around an hour to create an image and deploy it, will report back once i tested the image :) [17:06] bumblefuzz: well, one is a regular file, and the other either is or isn't - there isn't much insight required there [17:07] well, i just wanted mild check [17:07] but yeah, moving; thanks anyway :) [17:10] bumblefuzz: ubuntu-mate-18.04.2-beta1-desktop-armhf+raspi-ext4.img <- cmdline.txt and config.txt are regular file here as well [17:10] bumblefuzz: where did you download that image from, and what's its checksum? [17:11] I torrented it [17:11] bumblefuzz: yes, but where did you get the torrent from? [17:11] the ubuntu-mate website [17:11] sha256sum: bb74b607da2f4d417851e006fadd5de1304f84681db9c4bd8f17ff1b1d410995 [17:12] that looks alright - how do you mount the partitions? [17:12] maybe not the best place to ask this but does anyone know if it's possible to sync and iphone's music library over the network with a music library stored on a linux server? [17:12] bumblefuzz: oh wait, I see the issue [17:13] bumblefuzz: yeah, for some reason the content in /boot on the *second* partition is weird [17:13] bumblefuzz: no, actually - wait [17:13] even on the writable partition [17:13] bleb: There are probably apps to do that for iPhone, but that is a topic for a different channel please. [17:13] bumblefuzz: no, actually, it's fine - the first partition is supposed to be mounted under /boot/firmware, and those symlinks then resolve correctly [17:13] the /boot directory is incorrect [17:14] so, when I insert the SD card, both partitions mount [17:14] bumblefuzz: so if you write that image to an SD card that you plug into the RPi, and then boot it, the symlinks should be all fine [17:14] where do I look for the file [17:14] bumblefuzz: insert where? [17:14] I'm trying to edit those files BEFORE I run the RPi [17:14] so I'm mounting those partitions on my laptop [17:15] and looking to edit those files [17:15] bumblefuzz: ah, then it's no surprise that things don't mount as expected - because your laptop (I guess you've got some automounting going on) doesn't mount things in the way they're supposed to be mounted on the running system [17:15] ahh [17:15] bumblefuzz: if you want to edit those files, mount the first partition, and edit the files there [17:16] the symlinks in the /boot directory expect that the first partition is mounted under /boot/firmware [17:16] yes, things are automounting [17:16] (you can take a look at /etc/fstab) [17:17] my /boot partition only has bcm files and and /overlays/ directory [17:17] not any of those files [17:17] bumblefuzz: it does here - http://ix.io/1OgO [17:19] I don't see cmdline.txt and config.txt in there... [17:19] well, then something with your image must be weird - especially since you get a correct SHA256 checksum [17:20] if i want to lock sftp users' home directory on a separate volume mounted at /sftp, can i accomplish this with chroot? [17:21] will a unique home dir be created for each new user at that mount point? [17:21] becool: That isn't the purpose of chroot [17:21] bumblefuzz: what is the output of `losetup`? [17:21] basing it off of this article: https://websiteforstudents.com/setup-retrictive-sftp-with-chroot-on-ubuntu-16-04-17-10-and-18-04/ [17:21] hey people, i added new user 'administrator' and i'm trying to mkdir for the ssh but i get permission denied. https://askubuntu.com/questions/163589/cannot-create-file-directory-in-home-directory - this is exactly his problem but i tried example below and nothing changed [17:22] i'm really new to ubuntu just want to set up my droplet [17:22] becool: yes, with sftp chroot you disallow users from cd-ing outside of it. [17:22] haiiokarin: what directory are you trying to make that under? try 'sudo mkdir nameofnewdirectory' [17:22] no, don't try sudo whatever [17:22] blackflow: thanks [17:22] i think that this is under root directory [17:22] don't blindly throw sudo if you don't understand the issue [17:22] haiiokarin, 'ls -l /home/' ? [17:22] that was previously created by root [17:23] haiiokarin: that's why, you need super user / root privileges to make something under the root folder [17:23] becool: is it good approach or not? [17:23] haiiokarin: how did you create your 'administrator' user? [17:23] ayekat: adduser administrator [17:23] haiiokarin: so did you go through all the questions, and create a home directory for that user? [17:24] ayekat: i went through all questions but didn't create directory for him [17:24] ah well, then it probably drops you into / after you log in - where you don't have permissions to create anything (for good reason) [17:24] haiiokarin, that is one of the questions, I believe. Why didn't you create a home directory? [17:25] haiiokarin: create a home directory for your user (check the output of `echo $HOME`) [17:25] in the sshd_configure, do i need to add a line 'ChrootDirectory /sftp' in order to route and lock sftp users to it? [17:25] ayekat: under root to create for him right? or during the setup of adduser? [17:26] haiiokarin: first, have you *set* a home directory for that user? [17:26] adduser will copy /etc/skel, which you probably want rather than an empty home [17:26] (something like /home/haiiokarin) [17:26] in other words, when creating a new user, how do i create their home directory on /sftp instead of the default /home/username? [17:26] (ah no, /home/administrator, I assume...) [17:27] ayekat: just type 'adduser' and press enter [17:27] ayekat: no i haven't , after i added user it stayed on the same path as before creating [17:27] you'll get an interactive prompt for the rest [17:27] just switched from root to admin [17:27] haiiokarin: ... alright, sanity check - what is the output of `getent passwd administrator`? [17:28] ayekat: administrator:x:1002:1002:,,,:/home/administrator:/bin/bash [17:28] that looks sane - a home directory is set for that user === SimonNL_Afk is now known as SimonNL [17:29] so I would propose that (as root) you create that directory, then `chown administrator:administrator /home/administrator` to give the 'administrator' user permissions to read/write in it [17:30] ayekat, haiiokarin alternatively, copy /etc/skel instead of making an empty directory [17:30] did you add 'administrator' to sudoers? [17:30] it is copied though [17:30] /etc/skel is a template for new users [17:30] using `visudo` [17:30] i see bunch of folders when ls -l [17:31] or yeah, if you somehow particularly care about skel, maybe copy that... [17:31] which is home among others [17:31] haiiokarin: you see that bunch of folders *where*? [17:31] haiiokarin, you are not in the right directory, then. [17:31] Having a .bashrc and .profile is nice... [17:32] can i post you YT that i'm following? Just in case as i'm curious how he as i followed every step does have permission to write under 'administrator' [17:32] i'll point to the minute when he is creating user [17:32] haiiokarin: no, just tell us what your system currently looks like [17:33] youtube guides are rarely a good resource for information [17:33] how do i add a user as authorized for a specific private ssh key? [17:33] becool, please clarify [17:33] becool: append them to the authorized_keys file? [17:34] ayekat: ok so i'm now logged in as 'administrator' in my main folder 'icecast-server' [17:34] ryuo: thanks i think that's what i need to do [17:34] ayekat: which previosuly was created by root [17:34] in the .ssh directory for their home directory [17:34] haiiokarin: yes, but what about your home directory? [17:34] becool: be warned, sshd is picky about file permissions of these. it has to be accessible by only the user or it won't even read it. [17:34] haiiokarin, please tell us what 'echo $PWD' and 'echo $HOME' output. [17:35] ayekat: i do see home directory in the icecast-server path, but was it previosuly created by root or by administrator now? [17:36] lordcirth: yeah one second [17:36] administrator@icecast-server:/$ echo $PWD [17:36] administrator@icecast-server:/$ echo $HOME [17:36] /home/administrator [17:36] administrator@icecast-server:/$ [17:36] sorry on $PWD it was '/' [17:36] ryuo: do i have to create a public key for each sftp user i add to the server? [17:37] haiiokarin, so, administrator's home is /home/administrator, as it should be, but you are not in it. [17:37] lordcirth: cd /home/administrator ? [17:37] becool: you can reuse keys, but each needs their own keys file. [17:37] haiiokarin, 'cd' or 'cd ~' or 'cd /home/administrator' will take you to your home directory. Can you write there? [17:38] becool: just how it is for ssh in general. [17:38] haiiokarin: first, what does `ls ~` show? [17:38] (because if `cd` fails, you'll remain in /, and we'll be none the wiser) [17:38] ryuo: if the .ssh directory and authorized_keys file doesn't exist in the users' home directory, should i just run `mkdir` and `touch` to create them? [17:38] ayekat: yes one second [17:38] ryuo: or do i need to generate a public key for each user? [17:39] becool: yes and yes [17:39] you can also put their authorized_keys files outside of their homes, eg somewhere under /etc/ so they don't mess with it [17:39] cd /etc [17:39] ls [17:39] ayekat: ls ~ does not show anything, but ls -l does show folders [17:39] oops wrong window [17:40] haiiokarin, 'ls -la ~' ? [17:40] haiiokarin: yes, ls -l is expected to show folders - probably in / still [17:40] becool: erm. normally the *client* generates their key pairs [17:40] and gives the public key portion to the server. [17:40] lordcirth: administrator@icecast-server:/$ ls -la ~ [17:40] total 24 [17:40] drwxr-xr-x 2 administrator administrator 4096 Jul 12 17:24 . [17:40] drwxr-xr-x 5 root root 4096 Jul 12 17:15 .. [17:40] -rw------- 1 administrator administrator 382 Jul 12 17:38 .bash_history [17:40] -rw-r--r-- 1 administrator administrator 220 Jul 12 17:15 .bash_logout [17:40] you can also reuse keys for multiple accounts. [17:40] ryuo: i know, but in this case i create a test user call 'sftp' and am trying to sftp to the server with that user to see if it works [17:41] i see. [17:41] !paste | haiiokarin [17:41] haiiokarin: For posting multi-line texts into the channel, please use https://paste.ubuntu.com | To post !screenshots use https://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [17:41] then try putting it in their home directory as previously stated. [17:41] haiiokarin, however, that looks correct. [17:41] it's the default place to look for them. [17:41] That did not solve my problem [17:42] https://paste.ubuntu.com/p/jMsHNshZnQ/ [17:42] just so you see it better [17:42] Mikjaer, you should mention whoever you are talking to [17:42] lordcirth: not talking to anyone in particulair [17:42] haiiokarin: if you `cd`, you will change to your home directory [17:43] haiiokarin: and then you can start creating files and whatnot, according to whatever online guide you are following there [17:43] Mikjaer: this being a rather busy chan, you might want to summarize your problem into a pastebin, with examples, and post the URL here when you re-ask for help. [17:43] ayekat: it's just configuration of icecast server so it's rare to find anywhere thus i follow this one [17:43] ayekat: by just typing cd he will go back one step into home? [17:44] ayekat: he :D [17:44] haiiokarin: `cd {path}` will switch to {path} -- if you don't specify {path}, it simply switches back to your home directory (i.e. /home/administrator) [17:44] also, ~ is a shorthand for your home directory [17:45] ayekat: let me try making a directory [17:45] I recommend searching the web to learn the basics of command line usage - controlling a server will otherwise be quite a pain, both for you and the people in here [17:46] ayekat: i will - directory is created here but for the future i'll need to learn more command line usage [17:46] ayekat: do you have any recommended material? [17:46] haiiokarin, https://linuxjourney.com/ [17:46] lordcirth: thank you :) [17:48] Am i right to asume that the kernel for ubuntu-desktop comes with more drivers compiled in than the server one? And is there any way to install the desktop kernel on a server install? [17:50] the kernels should be the same for both - but additional driver packages may not be shipped by default with the server version [17:50] Mikjaer: no, it's the same linux-image-generic kernel and dependency packages with extra modules. [17:50] The r8169 does not seems to be loaded as a module on the ubuntu 18.04 desktop [17:51] but on the server version it is ... and it's not working [17:51] Mikjaer: it's loaded here in my case [17:51] Bakso: on a desktop? [17:51] ah yes ... there i see it, sorry [17:51] Mikjaer: yes, 18.04.2 [17:52] blackflow: no .1? [17:52] .2 is latest [17:52] blackflow: and? [17:52] and what? [17:52] why is that important? [17:52] Mikjaer, why are you asking about .1? I am confused [17:53] Mikjaer: because it's different from 18.04.1 [17:53] Mikjaer: if you install it from scratch, you'd get some different package choices, eg. it installs the 4.18 HWE kernel, nvidia-driver, automatically === SimonNL is now known as SimonNL_Afk [17:53] blackflow: "it" being? [17:54] Ubuntu installer [17:54] Yeah, if you install 18.04.0, you will not automatically get HWE when upgrading [17:54] Okay så they added better hardware support in .2? [17:55] HWE kernel is 4.18 [17:55] still doesnt explain why it works in desktop and not in server [17:55] Mikjaer: one could say that, by virtue of the HWE kernel being available. [17:55] good morning [17:55] Mikjaer: it? [17:55] psilly0: www.nohello.com [17:55] Mikjaer, what is your kernel version on the server? IIRC you said the desktop was 4.15? [17:56] blackflow: the network card :) [17:56] Mikjaer: it's the same model in both machines? [17:56] Mikjaer: can you pastebin, for both, the output of udevadm info /sys/class/net/enp3s0 or whatever the NIC name is instead of enp3s0 [17:56] blackflow: yes, and the machines are 100% identical [17:56] but i think i will try to download .2 [17:57] Mikjaer: you don't have to reinstall, you can install the HWE kernel and uninstall the LTS one [17:57] if that makes so much difference, i just didn't expect a minor version to do anything that normal packages didn't get me [17:57] blackflow: im working on a disk-image thats going to be deployed on a lot of systems, i'd rather go with a clean one for reproducability :) [17:58] sudo modprobe -r r8169 && sudo modprobe -i r8169 # should fix it [17:58] Mikjaer: it's a point release. Really no sense in requiring a reinstall to upgrade from .1 [17:58] I am curious as to if there is any documented features that are equivalent to Intel's Management Engine, or AMD's Platform Security Processor present in the newest Raspberry Pi 4, Or It's Broadcom BCM2711 CPU? [17:59] blackflow: and by the way ... i acnnont post the output of these commands, unless you want me to take a picture of the screen :P [17:59] iliketurtles99a8: wrong channel for that though. [17:59] blackflow: and by the way ... i can not post the output of these commands, unless you want me to take a picture of the screen :P [17:59] Mikjaer: because no network? [17:59] yup [17:59] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1752772 [17:59] Mikjaer: well ... can you spot any differences? [17:59] Launchpad bug 1752772 in Linux "r8169 ethernet card don't work after returning from suspension" [Undecided,New] [18:00] blackflow: im doing the install in a virtualbox, exporting the images and writing the image to a harddrive, and moving it to the physical hardware. Might that be part of the problem? Maybe some kind of probe is being run during install that i need to rerun myself? [18:00] ... [18:00] really.. [18:00] Mikjaer, why do you no ttell such info at the beginning? [18:01] Mikjaer: in general, there should be no problem with that, except you'd need to adjust UUIDs for mountpoints, and NIC names [18:01] however, there's no r8169 under vbox [18:01] vbox either has the virtio driver, or emulates Intel Pro something something 1Gbps something [18:01] without proper info, you get funny answers [18:01] e1000e driver iirc [18:02] which is _emulation_ so there might be issues. there have been issues. [18:02] blackflow: i know, the problem arrises once i move the images to physical hardware [18:02] OerHeks: because it's not essential [18:03] OerHeks: no need to confuse people with unneccesary details :) [18:03] Mikjaer: this info kinda is essential though. not all people know the specifics of virtualized hardware. [18:04] Mikjaer: for example the NIC names change. udev sees different hardware. but anyway, rebooting on the actual hardware should reinit hardware detection. [18:04] blackflow: yea i know that, and thats not the source of my problem :) [18:04] Mikjaer, well you got your answer, blame virtualbox [18:05] OerHeks: i can hear that you did not understand my question, but thanks for trying :) [18:05] Mikjaer: comparing those udevadm info outputs would be a great way to start troubleshooting. [18:06] what's the command to connect to an sftp server? [18:06] blackflow: execept from adresses begin different and USEC_INITIALIZED being different, they are the same [18:06] how can i list files in the home dir? [18:06] becool: sftp username@hostname [18:07] Mikjaer: dmesg is the next step. it will log driver autodetection and possibly loading failure. could be firmware, methinks r8169 needs firmware [18:07] Mikjaer: how do i specify a public key to connect with? [18:07] i dont se any firmware issues [18:07] becool: -i [18:07] Mikjaer: tried dmesg | grep r8169 ? [18:08] let me just boot the broken one again [18:08] Here's mine: https://dpaste.de/tOqV/raw [18:09] works! [18:09] btw, you specify the _private_ key on the client side. but I guess you figured that out. [18:09] now that i'm connected, i see that i am not in my own home directory, but the root directory specified in the sshd_config [18:09] How do I check if my network interface/driver supports promiscuous mode? [18:10] how do i make it so i sftp into the users' home dir? [18:10] also, i want all user's home dirs to exist on the sftp root mount [18:10] becool: pastebin your sshd_config please [18:11] sure one sec [18:11] blackflow: no 8169, but theres a 8168 [18:11] becool: you specify home dirs when creating users, with -d or -b [18:12] omg i'm an idiot, i just deleted my private key on accident and can't ssh to the server [18:12] anyway to recover the deleted file? [18:12] gesundheit! [18:13] becool: if you didn't disable it, chances are you can log in with password. [18:13] blackflow: okay ... the non working install (server) tries to use the r8168 driver, while the working one uses r8169 [18:13] it's an ec2 instance so it's cert only [18:13] Many users dont' know they must explicitly disable passwords, or else pubkey offers no increased security. [18:13] i should be able to issue a new key via aws console [18:13] Mikjaer: for exactly the same NIC hardware, and exactly the same kernels? [18:14] becool: tried ssh-ing in with password? [18:14] blackflow: dekstop : Ubuntu 18.04.1 LTS (GNU/Linux 4.15.0-46-generic x86_64) [18:14] there is not password [18:14] never defined one [18:14] becool: "root" user? [18:14] becool: try lsof -n | grep deleted -- if you're really lucky another process has the file open even though it's deleted. [18:14] amazon doesn't allow anyone root user access [18:14] sarnold: and then? :) [18:15] blackflow: server: 4.15.0-54-generic #58-Ubuntu [18:15] becool: then what does it do? creates a non-root user, with no password, and a predefined pubkey? I doubt that [18:15] Mikjaer: try upgrading to latest, could be bugs fixed that show this difference [18:17] i just tried rmod'ing r8168 end modprobe r8169, changed nothing. [18:17] blackflow: would an apt-get dist-upgrade get me to .2? [18:17] Mikjaer: yup, or "apt upgrade" [18:18] just tried en apt upgrade before i even asked inhere [18:18] Mikjaer: but the point here is not .2, but kernel version. `apt upgrade` should bring you to latest package versions state, on both. [18:20] just checked, it's upgraded [18:21] Mikjaer: so both kernels are at the same version? [18:21] hey folks, i would appreciate some help. anything i try to download i get the following message. "E: Unable to locate package wget" [18:21] blackflow: no, the non working one is fully upgraded [18:21] i use apt-get install wget using sudo [18:22] how can i install nano & wget [18:22] gavimobile: **how** do you try to download anything? [18:22] hello experts [18:23] guntbert: please see my answer above [18:23] Hi, I had a power outage while watching a video with vlc and now the fullscreen popup controls don't pop up. I tried reinstalling ubuntu, upgrading to current from lts and deleting ~/.cache/vlc and vlc preferences. [18:23] Mikjaer: not sure what to do next, if you don't see any meaningful errors about why the module didn't succeed. you can, however, fix this particular machine by forcing the driver in /etc/modules so it gets included on boot. [18:23] marian090909: www.nohello.com [18:23] anybody know why wine crashes whenever i try to install anything in playonlinux? is it not sustained anymore? [18:23] gavimobile: weird, apt/apt-get should not require wget (I think...) [18:24] blackflow: im thinking that i might just use ubuntu desktop, and then remove the gui [18:25] Mikjaer: as far as the low level stuff is concerned, which is about kernels and drivers, udev and plumbing etc..., it's the same. the difference between desktop and server is only in dependencies pulled in by metapackages. ubuntu-server vs ubuntu-desktop, while both have ubuntu-minimal and ubuntu-standard; and of course what those higher level packages do differently (eg. NM vs systemd-networkd on [18:25] desktop, for network management) [18:26] blackflow: i agree ... thats why it's weirding me out [18:27] gavimobile: Try it with sudo apt install wget (instead of apt-get) - I don't really expect success, tbh [18:29] Mikjaer: had to cut my losses and create a new image [18:32] gavimobile: ouch, I completely misread your question - sorry! You should start with the command sudo apt update to update the databes. When this gives errors tell us about them. [18:32] *database [18:33] can someone give/link me a rundown of the 32 bit story of the past week or two? [18:33] sarnold: btw, would you happen to know what's the state of ZFS support on Ubuntu's 5.x kernels? are those GPL only changes from mainline patched out? Or does ZFS take a perf plunge with 5.x on Ubuntu [18:33] blackflow: I'm assuming the performance plunge [18:34] sarnold: gah! [18:34] I should run me own benchmarks to see how much... [18:38] The OpenJFX package doesn't seem to work. I've tried various guides online, and none of them works for me, including exporting ENV and whatnot. Has anyone gotten OpenJFX to actually work so that Java applications that rely on JavaFX will run, and can be compiled? Thanks. [18:45] blackflow: just made a new install based on .2 and changed nothing after the installer, same problem. [18:45] TomyLobo: i'm not aware of a story during th past two weeks, but there is https://bryanquigley.com/pages/papers/ubuntu-drop-i386.html [18:46] i will try to install directly on the machine, to se if its something during the installer im missing out on [18:47] TomyLobo: this channel is just about support, though, there's also #ubuntu-discuss [18:48] Mikjaer: can you modprobe it? [18:48] blackflow: no, tried that before [18:49] Mikjaer: anything in dmesg when you try to modprobe it? [18:49] no, doing the excact same thing === SimonNL_Afk is now known as SimonNL [18:50] it just behaves as theres no cable in the nic [18:51] Mikjaer: do you have linux-firmware package installed? [18:51] yes [18:51] Mikjaer: $1M question... .you sure the NIC is not broken? :) [18:51] tried to swap the hardware? [18:52] blackflow: several times, i switches the drives 20 times before i fetched another machine so that i could have one of each [18:53] Mikjaer: but did you try to take the NIC out of that machine, and put it into the machine where that same model works? [18:54] Mikjaer: or I suppose those are onboard NICs? NO separate cards? [18:54] blackflow: onboard nicks, and two identical machines [18:54] blackflow: and the nics works if i boote the machines on the ubuntu-desktop-harddrive [18:55] blackflow: and here comes the funny part ... i just booted the system with the exact same iso that i used to create the ubuntu-server-image ... and the installer gets an ip adress on the nic [18:56] blackflow: so _something_ is happending during install that does something to it [18:57] Mikjaer: I doubt it. [18:58] The only difference is that i ran the installer on a different machine and moved the disk [18:58] Mikjaer: Ubuntu installer works with images that it unpacks. Other than partitioning, there's no special configuration it does . [19:00] blackflow: after installation, during first reboot, it hangs while trying to up' the network [19:00] some packages are automatically removed after the install is complete [19:02] Mikjaer: the part I'm failing to understand is what exactly is the symptom there. you can't modprobe? no error explanation? you can, but no network? is the config correct? [19:03] blackflow: no, as i explained several times. The NIC acts as theres no cable plugged in. [19:03] Modprobe works fine, device shows up, but state is down [19:04] Mikjaer, can you paste lspci -nnk ? [19:04] ioria: sorry [19:04] i can take a picture of it :P [19:04] Mikjaer: state will be DOWN until you configure it [19:04] Mikjaer, i see [19:05] okay ... the manual installation came op [19:05] blackflow: no it wont, it will change to UP when theres a link [19:06] Mikjaer, you see 'NO-CARRIER' in the 'ip -a' output ? [19:06] Mikjaer: did you _try_ configuring the nic? [19:06] also, I was convinced all this time the problem was what you originally stated: "19:50 < Mikjaer> The r8169 does not seems to be loaded as a module on the ubuntu 18.04 desktop" [19:06] ioria: let me just rewrite the image, and i'll check [19:06] Bakso: same to you [19:07] bl [19:07] Mikjaer, sorry... 'ip a' (without the -) [19:09] blackflow: im pretty sure i tried fetching an ip address [19:09] hi, I have a Raspberry Pi zero W that is ARM v6 [19:09] which ubuntu version will work with this? [19:09] !pi | bumblefuzz [19:09] bumblefuzz: Instructions for installing Ubuntu onto Raspberry Pi systems can be found at https://wiki.ubuntu.com/ARM/RaspberryPi [19:10] blackflow: but im uncertain ... and this whole thing makes so little sense, that it would make sense that i have missed something along the way [19:11] Mikjaer: thing is you installed in vbox. by default ubuntu installs netplan which configures the NIC name in its config. If you move that to baremetal, you'd have to change the netplan config, as I originally mentioned (UUIDs for mountpoints and NIC names). Did you do that? Or is netplan trying to configure for the no-longer present vbox NIC? [19:11] bumblefuzz: iirc ubuntu requires minimum armv7. I don't think ubuntu will work for you; give debian a look [19:12] blackflow: no i did not, i checked the interface with "ip a" and saw that it was down. [19:13] Mikjaer: well then... :) again, configure it. It's down until it's configured. === RandolfR is now known as Randolf [19:13] blackflow: so what your saying is that i should have done (the equivalent of) ifconfig up , and it would have worked? [19:14] Mikjaer: the equivalent yes, as ifconfig is deprecated. not just "up" but actually give it an address or use dhcp [19:15] Mikjaer: or simply modify /etc/netplan/... default config, put in the correct NIC name and run `netplan apply`. Let us know if that fixed it. [19:15] blackflow: but if thats the problem, then ubuntu-desktop must be doing something, to autodetect those settings [19:15] Mikjaer, usually server edition screams like hell if it cannot configure the nic [19:16] Mikjaer: ubuntu-desktop is using NetworkManager which yes, autodetects on boot. for servers, NIC name is baked into netplan config. [19:16] That makes sense :) [19:17] That was actually my first question, but i guess that was before you joined the conversation :P [19:18] Mikjaer: I'm sorry, I thought you'd pick it up from me stating you should change UUIDs and NIC names when moving from vbox to baremetal :) You never asked what I meant, so I assumed you understood. [19:18] Mikjaer, 18.04 configures netplan both on desktop or server; but on server it sets networkd on desktop NetworkManager [19:18] blackflow: i did ... i just thougt that the state in "ip a" was independant of configuration [19:18] ioria: yes and with that bakes in the NIC name, unlike on desktop that only configures NM as backend [19:19] Mikjaer: it never is. bfore netplan, it was ifupdown that did it. It also baked in the NIC name into /etc/network/interfaces [19:20] so the problem here is the interface name ? [19:21] blackflow: yea ... i can "up" the interface manually [19:22] Mikjaer: please try with netplan config and `netplan apply` [19:22] ioria: yeah, that's what I'm thinking, since this installation was done under vbox and then moved to baremetal. [19:22] ah, ok (vbox, i'am out) [19:24] blackflow: yea, that worked. Seemed the problem was me now knowing how the "new" ip command works :P [19:25] Hello! was wondering if I can get some help. My Ubuntu is hard crashing (only way out is to turn off my computer) whenever I watch any videos on Twitch (in browser + GNOME Twitch). Can anyone help me out in pinpointing the exact problem to avoid this? Would appreciate it [19:25] blackflow: Guess i owe you a beer, thanks :) [19:25] Mikjaer: awesome :) [19:26] ndeeah: next time try : Alt+SysRq+b , that should reboot your machine instantly. [19:26] Mikjaer: alrighty will do. any idea on how I can avoid it though? [19:26] ndeeah: no [19:26] ndeeah: need to know whats wrong before we can fix it. [19:27] Mikjaer: alrighty, got any pointers on where to look to identify what's going on then? [19:27] Not sure which log to look at [19:28] ndeeah: i just gave you one, if that works, then your kernel is not crashed, and it's probably and x-related problem. [19:28] Mikjaer: oh okay. let me open twitch & crash it then :D is SysRq the print screen button? [19:28] ndeeah: on some keyboards yea [19:28] You can also try "r" først [19:28] You can also try "r" first* [19:28] and then ctrl+alt+1 [19:29] that should give you a terminal [19:29] if it does, then your x is crashing [19:29] could be a driver-issue or maybe a dead gfx [19:30] Mikjaer: alrighty on it. Quick question, my /proc/sys/kernel/sysrq says 176. Should that be a problem? (assuming it should say 1) [19:30] ndeeah: look into /var/log/Xorg.1.log after you reboot, as well as journalctl -n 100 -b -1 for last 100 logged entries before reboot [19:30] ndeeah: mine says 176 to :) [19:30] Alright on it [19:31] it's a bitmask. 176 = 128 + 32 + 16 . /etc/sysctl.d/10-magic-sysrq.conf explains the numbers [19:33] Mikjaer: just another quick check. ALT+sysrq+r and then ctrl+alt+1? [19:33] Trying it now before even getting a crash, not working for me [19:35] * solderfumes sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/OnEbaPTnKQmnMijUEuaXjarA > [19:36] ndeeah: try it with b? [19:36] solderfumes: can we help you with something? [19:36] (that will reboot your machine) [19:36] solderfumes: is there anything in dmesg? [19:36] ndeeah: i think Mikjaer meant ctrl-alt-F1 [19:36] everything works, apart from the do-release-upgrade script [19:37] atrus: your right! [19:37] ndeeah: i ment F1 [19:37] although F1 is usually where the graphical login manager is these days. might try ctrl-alt-f3 or f4 [19:38] the file /etc/udev/rules.d/70-persistent-net.rules seems to have been moved, where is it not? I guess its a systemd thing? [19:38] look at /lib/udev/rules.d [19:40] Mikjaer: great, can't get it to crash now haha. The video is lagging like crazy though [19:41] solderfumes: i cant find the nic-definitions there though === Kartha is now known as { === { is now known as {-_-} [19:45] Mikjaer: i wonder if that's been replaced by netplan stuff. i have a mac address for my interface in /etc/netplan/ [19:49] i seem to have gotten downright good at crashing the compositor. when it comes back after a good 20s freeze, anti aliasing is in a stat of flux-freakout. has anyone seen this in ubuntu 19.04 ? nvidia drivers 430 [19:49] in a dual boot scenario, can Ubuntu's EFI partition be on the same disk as Windows'? can/must they share the same EFI partition? (when there's only one disk) [19:51] Mikjaer: (yes, i believe this is handled by netplan these days: https://serverfault.com/a/941659 ) === {-_-} is now known as [0_0] [19:52] When running `do-release-upgrade`, the script will download a tarball (disco.tar.gz) and a signature (disco.tar.gz.gpg). Can anyone tell me what key is used for signing this tar file? Wherever does it download it from? === [0_0] is now known as {o_o} [19:53] When running `do-release-upgrade` the script won't authenticate the downloaded files, but there is little info available === {o_o} is now known as Guest58455 === Guest58455 is now known as tens0r === tens0r is now known as hippopotoneus === Gokturk-Away is now known as justme23 [19:59] solderfumes: do-release-upgrade (and its GUI equivalent) connect to https://changelogs.ubuntu.com/ and download the relevant meta-release-* file based on LTS and "upgrade to development releases" flags and settings. that file contains the URL of the upgrade tool (the .tar.gz you mentioned) and its gpg signature [19:59] solderfumes: i don't have a gpg install handy, but given the .gpg file it should be able to tell you the key it was signed with [19:59] i assume it's one of the ubuntu release keys [20:00] solderfumes: /etc/update-manager/meta-release -> UR{,_LTS} -> UpgradeTool{,Signature} [20:00] solderfumes: /etc/update-manager/meta-release -> URI{,_LTS} -> UpgradeTool{,Signature} (typo corrected) [20:03] i found the file, it was signed by 3B4FE6ACC0B21F32 and/or 871920D1991BC93C, let me check if these keys are legit [20:05] I found the key, it's on my computer at /usr/share/keyrings/ubuntu-archive-keyring.gpg === justme23 is now known as Gokturk-Away [20:05] should this key be in /etc/apt/trusted.gpg? [20:08] that did it [20:09] thanks for the help, dax, TJ-! [20:11] getting ' Unable to mkstemp /tmp/clearsigned.message.gMZ8z8 - GetTempFile' [20:11] I am currently checking google, but hoped someone here might have an idea wft? [20:12] solderfumes: if you need to know which key signed a file (whether inline or detached), and have the file containing the signature: "gpg --verify bionic.tar.gz.gpg /dev/null" [20:13] autoclean... autoclean was the answer. Thanks and sorry for wasting your time folks :D [20:15] TJ-: I didn't know about the /dev/null trick, I just gave it the tar.gz file, it worked just as well. Adding this signing key to my /etc/apt/trusted.gpg allowed me to run `do-release-upgrade` [20:17] solderfumes: strange; the keys should all be in the existing ubuntu-keyrings and by extension, available to apt-key [20:17] solderfumes,TJ- -- be careful with that though, you don't want to just blindly use whatever key an attacker-provided blob says to use to verify it :) [20:18] sarnold: well of course; the point is to get the KEY ID of the signature [20:19] just don't forget the next step -- verify that that keyid is a legitimate key id :) [20:19] who cares? it's only bits. Easily clobbered with the off switch :) [20:21] i cannot see the "VPN Connection" in the top right menu on 19.04 with ubuntu desktop, any hints? [20:21] *I* care :) we've gone to great lengths to provide a chain of trust from our build servers to our users and if users just blindly grab whatever keys were used to sign something, that breaks the chain :) [20:24] sarnold: that's a good suggestion, but the key was already on my machine, just in a different location [20:25] solderfumes: yeah, I'm glad you looked for it there and found it :) I'm surprised it wasn't already there, but.. you got the right thing. [20:25] TJ-: I agree, it's strange. This machine was distro-upgraded a few times, and maybe that's the reason the archive key was not in the apt/trusted.gpg [20:26] solderfumes: if you "dpkg -L ubuntu-keyring" you should see it is listing keys in /etc/apt/trusted.gpg.d/ [20:26] (gpg command line should be rewritten from scratch IMO, it is an absolute pain with keyrings which aren't in your .gnupg directory) [20:27] TJ-: that's exactly what I did :) [20:27] then the usual `gpg --no-default-keyring --keyring /usr/.../full-freaking-path.gpg --export` etcetera [20:28] solderfumes: and while we're at it, maybe we can move to crypto from *this* decade, too.. a complete replacement would be beautiful. [20:29] Is there any way to _completely_ disable Gnome notifications? [20:29] can I uninstall a Gnome component or something like that? [20:29] sarnold: What advancements have there been in cryptography since the past decade? What is your suggestion? [20:30] sarnold: I wish the same care was applied to releases.ubuntu.com (no https) so the images and signatures there could be trusted not to MITM-ed - this for new users coming fresh to Ubuntu and not having skills/access to gnupg/PGP on their current OS. Try switching to HTTPS and it delivers a certificate for "*.bit.nl" [20:30] solderfumes: authenticated encryption modes as the only way of operation, and much less of postel's law "liberal in what you accept".. [20:31] TJ-: yeah, users from windows have basically no options short of "use wsl to install linux first and then use that" [20:31] dconf org/gnome/desktop/notifications [20:32] cross-post from #systemd: How can I make systemd stop bluetooth more quickly on Ubuntu xenial? "22:15:17 bluetooth.target: Unit not needed anymore. Stopping.¶ 22:15:17 Stopped target Bluetooth." … then for almost 2 minutes, nothing happens … "22:17:11 Stopping Bluetooth service..." my problem is that when I touch the antenna the wrong way, it moves just enough to disconnect USB. then I can't just push it back, I have to res [20:32] tart bluetoothd if I want it to work again. (or probably wait 2 minutes, haven't tried that.) [20:34] OerHeks, was already switched off, Gnome does not seem to care about it though, spamming me with warnings that my mouse battery is empty. well its not and it is very annoying to get that popup spam every few seconds [20:34] oh, didn't you had this issue before? [20:34] level down the treshold? [20:35] yeah again and again, threshold set made it less frequently, but still now its starting again (didnt change threshold though) [20:35] i would replace batteries [20:36] it doesnt solve the problem, just delays when its starting with popup spam [20:36] Sven_vB: I'd presume if the underlying Bluetooth device has disappeared the kernel is left waiting for it to respond and eventually decides it isn't going to [20:36] org.gnome.settings-daemon.plugins.power [20:38] already have all the percentage stuff at 1 [20:38] de-facto: are the reports inaccurate or just premature? [20:39] I suspect inaccurate (how can i verify?) and sometimes extremely frequently (annoying me every few seconds). I suspect there is a bug involved somewhere since i cant imagine thats how the devs want to annoy their users [20:40] i really would like to completely disable this gnome component alltogether, i dont need notifications anyhow [20:42] https://gitlab.gnome.org/GNOME/gnome-settings-daemon/issues/108 [20:42] de-facto: I know that some time ago I noticed the battery for my Asus T300CHI's Bluetooth keyboard/touchpad dock was being reported accurately but at some point something broke it and it now always reports 99% .. I don't use Gnome but I'm wondering if we're both seeing symptoms of a bug/regression in the power/input devices logic [20:42] TJ-, but it instantly says it doesn't need BT anymore? [20:42] GNOME 3.32 fixed it [20:43] and could well be this '1.2 V instead of 1.5 volts (NiMH) thingy ? [20:44] Sven_vB: how do you mean? "it says2 "it doesn't" - what are these "it"s ? [20:44] OerHeks, yes its got a NiMH rechargeable AA which i can charge over USB [20:45] TJ- yeah it may be that its inaccurate, but i would consider such frequent popup spam a bug in itself already. sometimes when it pops up the mouse movement even stutters (i guess it does something with the logitec interface) [20:45] de-facto: you can check the underlying kernel view of the battery state with " grep . /sys/class/power_supply/hid*/capacity " [20:46] de-facto: I agree entirely about the notifications.... but I wonder if that could be because the reported level is changing frequently so it is say 99% one moment then 8% the next. I could imagine that causing repeated notifications and it would be no fault of Gnome [20:47] interesting idea, might make sense if that is the case [20:48] de-facto: you could put a watch on those values to see if they are changing with "watch -n 1 grep . /sys/class/power_supply/hid*/{capacity,status} " [20:48] i dont have capacity there [20:48] de-facto: really? what nodes are there? [20:49] de-facto: I wonder if that could be part of the issue? If something is expecting 'capacity' and it doesn't exist the code is triggering the notification [20:50] TJ-, https://pastebin.ubuntu.com/p/NwQTfmZfhy/ [20:50] ubuntu 18.04 amd64 [20:51] some older Logitech Performance Mouse MX [20:54] de-facto: "capacity_level" maybe ? [20:54] one with that logitec unifying usb receiver or what its called like (over their proprietary rf), the mouse containing a AA NiMH rechargeable 1.2V 2000mAh at currently 1.23V [20:57] the /sys/class/power_supply/hidpp_battery_0/capacity_level is "Critical" [20:58] de-facto: I'm reading a patch-set from 2016 for this hidpp_battery, and there it says that some of those devices do not report their charge level and some of them even report it as 0 when plugged in (that for solar-charged devices!) [20:59] yeah there is a keyboard from logitech that has solar panels inside [21:00] I would love if my system would either completely ignore any of those power levels or do more accurate warnings (no more than once an hour or something like that) [21:00] not every few seconds when something goes berserk about those power levels [21:01] its also not reliable, sometimes its very often, other times I cant remember it annoying me so often [21:01] de-facto: apparently there is/was a bug in UPower when dealing with these Logitech devices due to the missing nodes, so that could be what you're seeing [21:02] de-facto: see https://bugs.freedesktop.org/show_bug.cgi?id=100359 [21:02] Freedesktop bug 100359 in general "Add support for "capacity_level" attribute" [Normal,Resolved: fixed] [21:04] GNOME 3.32 fixed it [21:05] can i somehow disable upower? [21:08] disabling/stopping it via systemctl does not have any effect, it still is acrive [21:08] de-facto: I'd assume it's Dbus activated [21:08] de-facto: remember that most things freedesktop/gnome use Dbus for IPC [21:09] i just deinstalled gnome-power-manager, maybe that helps somehow... [21:10] maybe a bit drastic, but the popup spam stopped for now... [21:11] :-) [21:13] its was especially annoying because the popup goes over all the applications menu's (z layer above everything) so i had to click it away every few seconds [21:13] thanks guys :) [21:15] have fun! [21:22] I guess clean wasnt the whole story. I am getting: dpkg: error: unable to create new file '/var/lib/dpkg/status-new': No space left on device [21:22] but df -h reports 83% for that device. [21:24] MWM: yay you're back :) [21:24] MWM: df -i [21:25] hey df -i reports 100% in use... thats not right [21:33] found a page that shows how to identify the "files that are consuming inodes" but 'sort: cannot create temporary file in '/tmp': No space left on device [21:33] ' [21:34] TJ-, sorry for the wait. when I unplug the antenna, systemd detects it immediately: "22:15:17 Stopped target Bluetooth." but then it takes almost 2 minutes until "22:17:11 Stopping Bluetooth service..." [21:35] i RTFM and am still missing something on being able to create an sftp user with a home dir on the sftp folder/mount defined in sshd_config [21:35] i'll pastebin my sshd_config [21:36] becool, what did you try, what did you expect, what did you observe instead? [21:38] Sven_vB: hows your enterprise doing sir [21:39] i was thinking about getting support [21:39] here's my sshd_config: https://pastebin.com/TS6A1eaU [21:40] i want to be able to create new users (for sftp) and have their home directory automatically created under /sftp/users [21:40] Sven_vB: that makes sense. As I said, bluetoothd (controlled by bluetooth.service) will timeout as the device disappeared and the kernel is not sure what to do about it [21:40] when i connect to sftp as the user, i want it to go directly to their home dir [21:41] abbie, I'm open for business. if it's something that can be dealt with in public, you can also ask in the respective channel and maybe I can help for free. [21:44] is there an openssh channel specifically for this topic? [21:44] TJ-, then I probably just didn't understand what you meant to say. because when I stop BT via systemctl stop bluetooth, it quits immediately. [21:46] becool, so the problem is that they can see stuff outside their home directory? [21:47] well, the problem is that they never actually connect directly to their home directory, just the global root directory specified in the sshd_config === guiverc2 is now known as guiverc [21:48] becool, try ChrootDirectory %h [21:49] becool, see also the "ChrootDirectory" chapter in man sshd_config [21:51] cool i'll try that now [22:01] Sven_vB: that worked! you are my here [22:01] hero* [22:01] becool, you're welcome :) [22:01] have a gret weekend [22:08] hi unfortunately, my system is still crashing/freezing [22:09] it has a lot to do when I do anything visual like full screen a video or open media [22:09] fullstack: have you been able to monitor temperatures? [22:09] TJ-, if I type 'sensors' I get "PCI Adapter 60C" , and two other devices.. amdgpu-pci (no results) and acpitz-virtual-0 "+30.0C" [22:10] I hear my fan and don't think its overheating [22:10] I think its amdgpu [22:10] is there a way to stay with Ubuntu 16.04 _____LTS_______ (LTS should mean something) and utilize the latest amdgpu? I have had problems with kernels 4.20 .17 and .19, but I didn't try 5.x [22:11] fullstack: are you able to use another PC to ssh in and have it run "dmesg -w" to capture the kernel log? And of course boot the system with "debug systemd.log_level=info" on the kernel command-line - this will attempt to capture any messages that don't make it into the log-files due to the freeze [22:12] Ok yes that is very good information. I can put that line "debug systemd.log_level=info" in my grub defaults, correct? [22:13] I'll try to ssh next time [22:14] fullstack: yes. [22:15] because there's nothing in the logs now, this should at least be an effort to isolate the issue. thanks again. will know in 24-48 hours:) [22:15] fullstack: on the other PC use "dmesg -w |& tee /tmp/dmesg-problem-pc.log" so you can both see the messages and capture them to a file [22:17] fullstack: actually, to be more complete "ssh other-pc dmesg -w |& tee /tmp/dmesg-problem-pc.log" [22:59] hi, new laptop, XPS 15 7590, with Nvidia GTX 1650. I connect an external display and ubuntu recognize it and everything but it just stays black [22:59] I tried installing the nvidia drivers, tried disabling the nvidia card. nothing helps [23:02] elichai2, I setup nvidia 1050 the other day and all I had to run was 'ubuntu-drivers autoinstall' [23:02] does 'ubuntu-drivers list' show anything? [23:02] yeah [23:02] it shows the device [23:02] can you run nvidia-settings or whatever? [23:03] nvidia-detector [23:05] ha. it returns none [23:05] even though nvidia-smi detects it [23:05] laptop.. does it have a FN + screen key, internal/external/both? [23:06] yes [23:06] that would be a hardware key, even if it is F8 [23:06] elichai2, I just ssh to my machine with the 1050 and nvidia-dector returns none also [23:06] I tested that [23:06] elichai2, and nvidia-smi is OK [23:06] i'm actually running a job on it [23:08] ok. something weird hdmi through C works [23:08] direct HDMI doesn't [23:12] I had to ALT-F1, sudo to root [23:13] and then run 'init 3' and 'init 5' between enabling features in /usr/share/x11/xorg.conf.d/10-nvidia.conf, a file I had to create [23:13] this was ubuntu 16.04. I am not familiar with why there was no xorg.conf and everything had to be under /usr/share/x11/xorg.conf.d, but whatever. [23:13] 'nvidia-settings' command failed to run because there was no xorg.conf [23:14] anyway I am sure there is a solution, did you find one? [23:15] xorg.conf(5) discusses the different locations and how they work together [23:15] fullstack: did you modify files under /usr/share/ or just add files? [23:16] I created a file called 10-nvidia.conf [23:16] and stole someone's config [23:17] fullstack: https://github.com/JackHack96/dell-xps-9570-ubuntu-respin/issues/83 [23:18] fullstack: alright, good. modifying files in /usr/share/x11/... probably would have busted package updates. as tomreyn points out, xorg.conf(5) has the list of paths tht are consulted, and something in /etc/ would be less likely to cause trouble, and easier to find again in the future if needed :) [23:19] is there any reason why it was changed? it breaks the very nice 'nvidia-settings' application [23:20] there's been a push for a decade or two to move defaults to /usr and then have admins configure just what they need to change on the machine in /etc [23:20] with the long-term goal of eventually perhaps having just a handful of small files in /etc that's special about the system in question [23:21] the /etc/resolv.conf change is really annoying, removing the symbol link and chattr +i is the first thing I do installing Ubuntu. So there's definitely different objectives going on in regards to just changing /etc [23:25] see the /ETC/RESOLV.CONF section in systemd-resolved(8) on this if you haven't [23:25] there's always new and exciting ways to configure things but they aren't exactly better. yes, old ways are wrong and somewhat broken, and could be better, but they work. And the "enhancements" risk bring new and bad user experiences. Every half decade or so somebody attempts to do something and it breaks the old ways. that's why I wanted to run a LTS "please just boot up so I can get my work done and pay my bills without having to do a [23:25] bunch of unneccessary devops for a week" [23:27] LTS mostly ensures that there will be no architectural changes within this release. they cannot and do not attempt to prevent that there are no such changes across LTS releases. but this is rather a discussion to be had in #ubuntu-discuss === coffeeguy is now known as random-guy === random-guy is now known as coffeeguy [23:40] Hello, I wanted to add a disk to an existing 2 disk raid 5, but I think i fucked up. The original two disks were not used fully, but the other partitions on them were deleted. On the new disk i wrongly created just one big partition and added it to the raid. Now I cannot resize the filesystem. What I did till now: "parted -a optimal -- /dev/sdd mkp [23:40] art primary 0% 100%", "mdadm /dev/md1 --add /dev/sdd1", "mdadm --grow --raid-devices=3 /dev/md1 --backup-file=/tmp/md0.bak". Is there any way to fix this without rebuilding the raid multiple times? [23:40] This seems weird: md1 : active raid5 sdd1[2] sdc3[0] sdb3[1] [23:40] 55, hidden sectors 2048, dos < 4.0 BootSector (0x80), FAT (1Y bit by descriptor); NTFS, sectors/track 63, sectors 15628048383, $MFT start cluster 786432, $MFTMirror start cluster 2, bytes/RecordSegment 2^(-1*246), clusters/index block 1, serial number XYZ [23:40] : 82d7944:7c7e1e12 name=kubuntu:1 level=5 disks=3 [23:40] Is doing "mdadm -f /dev/md1p1 /dev/sdd1" and "mdadm -r /dev/md1p1 /dev/sdd1" and to readd the disk with a partition as big as the other ones the "right" thing to do? [23:41] thanks for all the hard work everyone puts into Ubuntu I appreciate it, and the help here. Its a wonderful OS [23:41] my statements were in general for any OS. If anything at all, Ubuntu gets it right better then any other OS [23:46] well im going to bed, but im thankful for any input