[00:00] heh, I made my /boot too small .. I did a snapshot of it after install, and then kept wondering why I had to keep cleaning up full /boot during every kernel update... [00:01] I have mirrored ssd's in my laptop, and I just didn't (don't) trust flash cells not to die. I know the stats say they are better than hdd, but old experience with flash is hard to erase. [00:01] sarnold: yes i was never brave enough to use zfs on /boot [00:01] Strontium: oh man, mirrored ssds on the laptop.. sounds good. I'd quite like to have eg four nvmes in this thing.. [00:02] this is an old laptop now, they are both sata. And i pulled the dvd drive out and replaced it with a 2TB hdd. The SSD's are only 128GB. [00:03] Its worked really nice. I have the os on the SSDs and a decent amount of L2ARC for the HDD which makes it feel a lot faster. [00:03] Its hard to find a laptop that can have enough drives to do this now. [00:05] Strontium: oh sweeeeet, that sounds like a nice setup [00:05] Strontium: I just had a Corsair SSD issue a SMART failure about a dying sector. [00:05] First time, of 4 SSDs and a few years, so I'm now joining your paranoid/dinosaur club :) [00:05] my old laptop had dual bays for drives, I had one ssd and one much larger hdd, and did rsnapshot I think between them.. [00:06] gbellinoz: yes, its a small club but i will always run ssd's mirrored if i care about the data. [00:06] The problem I always have with mirroring is it's 2x the hardware usage for the once-in-five-years failure day. [00:07] I've always just relied on hourly snapshots and off-site backups. [00:07] gbellinoz: yes, but thats offset in my mind by the improved performance. === Wulf4 is now known as Wulf [00:08] Do you ever mismatch disk sizes, or just always buy them in pairs? [00:12] gbellinoz: you do get increased read iops with mirrors though.. === thegoose is now known as el [00:37] Greetings! I dual boot Ubuntu and windows 10 on my laptop. The laptop has a spot for a microsd card. What should I format the card for both OS to be able to read from it. I plan to store linux distros and movies on it. [00:40] fat is the only filesystem both understand natively; ntfs is probably a good choice, even though the linux support for that is via fuse I think; exfat may also be a choice, but it's newer and perhaps not as reliable, and also done via fuse [00:40] ok thanks! [00:44] what happened to vfat? [00:45] ScottyK: Be careful. I had a windows install blow my Linux partition away one time out of probably 8 times doing that exact thing... so I was forced to give Windows its own SSD and unplug the linux SSD when installing Windows. Pretty shitty. [00:45] I would never format a micro SD card as NTFS [00:45] or FAT, really [00:45] Use exFAT, especially if you plan to use it with a camera [00:48] i've just installed ubuntu 16.04 LTS onto a laptop (i tried 19 but for many reasons, it didn't work well for me), and I need 16 because it's one I know I can install java 1.7 which is required for work. [00:48] odd thing: the virtual consoles don't appear when i hit ctrl-alt-Fn. they're THERE, because if i enter loging informaiton and run sudo reboot, the machine reboots. [00:48] but they don't appear on the monitor. [00:48] anyone seen this before? [00:48] i've googled lots of combinations of descriptions of the issue, but nothing. [00:50] dbtid: is the screen black or off or..? [00:51] the X screens are still shown [00:51] dbtid: Are they all missing? On 19.*, I was missing ctl-alt-F2 and F3, but F4 worked, and after I went to F4, sometimes 2 and 3 would work. [00:51] the mouse cursor disappers [00:51] dbtid: I never figured out why that was, though. [00:52] well, i've tried 1-6, no luck [00:52] the cursor in my text terminal disappears, and when i hit ctrl-alt-f7 i've got access to my GUI session [00:52] this is so weird [00:52] i edited /etc/systemd/login.conf [00:53] Ubuntu is definitely weird about the virtual consoles. [00:53] uncommented NAutoVTs and ReserveVT and rebooted, but that changed nothing [00:53] yeah, that didn't make any difference for me either. [00:53] well, when this laptop ran Ubu 14, all was well. [00:53] and it worked on 19 [00:53] 19 was making me nuts though [00:53] but 16 doesn't work [00:53] I'm guessing it might be hard to get much help for 16, that's a long time ago. [00:53] (banghead) [00:53] well, it's LTS :) [00:55] oh, and i tried running chvt, and i get the same effect as hitting the corresponding ctrl-alt-fn key :) [00:55] the keyboard is rerouted to the terminal, but the screen doesn't change [00:56] dbtid: so a few guesses .. (a) maybe it's nvidia-ish mess? that's way out of my experience.. [00:56] it's a radeon controller [00:56] dbtid: (b) maybe you can fix it by fiddling aruond with eg vga= or video= kernel command line parameters, or perhaps others.. they are documented at https://www.kernel.org/doc/html/latest/admin-guide/kernel-parameters.html but it's hard to know which ones to suggest [00:58] do-release-upgrade says I can't upgrade until i've installed all updates. But I've done apt-get dist-upgrade, upgraded everything, and even rebooted [00:58] I think this is because I have a held package. Well, I want the package held... [00:58] Is this why? How do I get it to upgrade regardless? [01:01] longus_catus, seems like a critical package. [01:02] gbellinoz: i bought these as pairs, but its not super critical, so 2 x 256GB SSD's (Different brand) would work fine. You may not get full utilisation of one if 256GB != 256GB but that shouldn't be a big loss. [01:02] https://ubuntu-mate.community/t/17-04-cant-switch-ttys-with-ctrl-alt-f-x/13043 [01:02] that exactly describes th eproblem [01:02] aha! the laptop is in the dock. [01:02] oerheks: It's firefox [01:03] firefox gets frequently updates. [01:03] gbellinoz: if i am not interested in data safety i would use ZFS and stripe a bunch of cheap small SSD's, you get incredible read/write performance doing this. Much better return than trying to get the fastest single SSD you can on the fastest bus. [01:03] Oh, you literally just mean "there is a package with an upgrade" [01:04] I guess I'll sed my /etc/apt and upgrade from there [01:05] longus_catus: If you hold back a package, anything that depends on that package will not update, as the dependency is because something is required. There is no way to force an upgrade, as what ever has the dependency will fail and thus other applications will begin to fail to run. You have to release the held package to get the update [01:05] I see what you mean. I forgot that the the version dependencies is both the pro and con. [01:06] so, a bit more, there's two video adapters: an intel in the laptop, and a radeon in the dock. [01:06] so i'll just undock to do what i want. [01:07] longus_catus: If firefox updating too frequently is a problem. i would recommend looking into Firefox Extended Support Releases (ESR) which only release when there are critical vulnerabilities being patched, and feature updates are far less frequent [01:07] dbtid: aha! I didn't think that happened with amd, I figured they were amd all the way through.. [01:07] dbtid: thanks :) [01:08] i got an ubuntu usb drive to work with unetbootin but my windows one didnt work, or didn't get read when I hit f12 [01:08] using mkusb now [01:08] So, you're saying that if firefox depends on a certain version of libc, libc can never upgrade while teh package is held? [01:08] my windows didn't work either [01:08] longus_catus: correct [01:09] Thanks for your specific suggestion about ESR [01:09] longus_catus: one possibility: install an ESR firefox from mozilla.org rather than from the distro, if they offer an ESR that's old enough. [01:10] hi trying to build termite, the build cannot find , installed libgtk-3-dev already. Any thought on how to get these header files? [01:10] longus_catus: While volunteers can't offer support for software from PPAs, this is where you can get ESRs through apt: https://launchpad.net/~mozillateam/+archive/ubuntu/ppa [01:10] snap install –channel=esr/stable firefox [01:11] :-) [01:11] henninb: is it a gtk-2 or a gtk-3 application? [01:11] or the snap is probably a safer option [01:11] = FF 60 [01:11] sarnold I am not sure if it is gtk-2 or 3. [01:12] to make a pop os bootable usb, with mkusb, what option do i choose the live version for ubuntu and debian one or the iso one on top? [01:12] henninb: alright, try installing libgtk2.0-dev and see if it gets further :) heh [01:12] sarnold, I will try [01:12] hi all [01:13] sarnold, libgtk2.0-dev is already installed. [01:13] same issue [01:14] hi [01:15] henninb: gtk programs typically need to be configured to find the location of the include files (which you can find with locate gtk/gtk.h) [01:15] So some of my users are coming from windows via VNC to ubuntu 18.04 and experiencing random freezes on the system running in VMware ESXI. I've checked top/htop/iostat and I'm not getting high IOWAIT, network latency is 2ms. Where else should I look? [01:15] henninb: what's the output of pkg-config --cflags gtk+-3.0 vte-2.91 [01:16] henninb: so maybe configure didn't work right -- look at the -I lines in the compile line that's failing, there should be something like -I/usr/include/gtk-2.0 [01:16] (or 3.0, whichever one it wants) [01:16] aernan: what are you using as a VNC service provider? [01:16] akk, I will see if I can find out how it is found in this autogen.sh. [01:16] akk: https://github.com/thestinger/termite/blob/master/Makefile [01:16] pragmaticenigma, default ubuntu stuff I believe. [01:16] henninb: There's probably a config.log which might be easier to search than autogen.sh [01:17] akk, will check [01:17] henninb: It'll still be long, but if you search for occurrences of gtk in config.log that might tell you something. [01:18] aernan: I would suggest you find out... there are a lot of options, and that's the only way a volunteer is going be able to help trouble shoot. There is no default ubuntu stuff for VNC servers... you could be running x11vnc, vino, or some other package. it's something you as the system admin would have had to make the decision to run [01:18] pragmaticenigma, I'm looking now [01:18] aernan: maybe check packet loss or error counters on the NICs in question? maybe a machine is dropping packets every now and then [01:19] akk, logs show this: configure:15450: error: Package requirements (glib-2.0 >= 2.40.0 gobject-2.0 pango >= 1.22.0 gtk+-3.0 >= 3.8.0 gobject-2.0 gio-2.0 gio-unix-2.0 zlib libpcre2-8 >= 10.21 gnutls >= 3.2.7) were not met [01:19] pragmaticenigma, vnc4server is the package name [01:19] sarnold, in a virtual environment will the vnic show actual packet drops? [01:20] henninb: I'm surprised it even generated a makefile if package requirements weren't met. But in the config.log you can see what it was actually looking for, so you can add whatever's missing. [01:20] akk, me too. usually it fails with the missage package. [01:21] sarnold, I do see a total of 33 drops over 6 days uptime. [01:23] with mkusb, for pop os i just choose the top selection , the iso one? [01:23] sarnold: https://bugs.launchpad.net/ubuntu/+source/ubuntu-release-upgrader/+bug/1850872 [01:23] Launchpad bug 1850872 in ubuntu-release-upgrader (Ubuntu) "When upgrading from 19.04 to 19.10 with ZFS Root Pool, Upgrader removed zfsutils-linux" [Undecided,New] [01:25] for pop os, i clone the iso with mkusb? [01:26] anyone? [01:26] bueller? [01:26] indoorcat_, not sure if this is a continuation from an earlier discussion... but popOS is not a supported flavor of Ubuntu for this channel, last I knew [01:29] aernan: and the hardware nics? that's probably not enough to be too worried... but check that number from time to time as people complain? [01:29] JimBuntu, nobody responded to my ppd issue on ask ubuntu [01:29] Strontium: thanks [01:29] man I just need tech friends to help me [01:30] indoorcat_, official doc appears to be https://support.system76.com/articles/live-disk/ [01:30] indoorcat_, I'm not surprised no one responded though, Canonical doesn't create popOS or control what they release/etc [01:31] Even though it's basically a skinned and modified Ubuntu, I don't think it's officially supported by Ubuntu any more than debian supports Ubuntu, so ##linux or their official channel for popOS would be the go to places, but we all like to help when we can [01:32] disks doesn't work for me, i need the computer to read it and i press f12 and then boot from usb [01:32] so far i can only make ubuntu usb bootables, windows and pop os doesn't work [01:34] If you boot from an Ubuntu live system, can you then make the popOS bootable? Either way, it could be something common between the popOS and Windows systems. I'm simply not up to date on what all has been changed from the official Ubuntu release to the popOS version of it to know [01:34] JimBuntu, I am using ubuntu 19.04 and having issues getting my printer to work, the usb bootables is just me having a work around so i can print tomorrow [01:35] I messed up my dual boot and am stuck with ubuntu [01:36] 19.04 is supported by the volunteers in this channel, not something I can help with ( too new ). Drop your `lsb_release -a`, printer make/model info and exactly what is happening and give it at least 10 minutes and you should hook someone, maybe longer, but give it some patience and someone should be around to help [01:36] g=brb === JanC_ is now known as JanC [02:42] I am installing a networked printer (standalone conencted via ethernet). I am sing the print manager and I can find the printer but it asks me what connection to use... Driverless IPP, AppSocket/HP JetDirect, or LPD network printer via DNS-SD. Which do I use? [02:45] Your printer setup page should have some information about what it even supports [02:46] or you could use nmap to see the open ports [02:47] longus_catus: nmap worked... excellent... thank you. Printer seems to support all three. Thank you. [02:49] Hi [02:49] somebody using dell 5490? [02:50] having troubles to recongize seagate usb drive of 1tb === zbenjamin is now known as Guest65914 === zbenjamin_ is now known as zbenjamin [02:56] acidtripper, use a paste service to give the output of lsusb [02:58] no output man [02:58] not recognized in kernel [02:58] Even if not recognized the usb subsystem should see the controller for the external drive [02:59] are you saying that lsusb gives you no output at all? [03:00] Also, if lsusb isn't seeing it, then it isn't showing up. Either it isn't getting enough power for the controller to be recognized or the drive is damaged. [03:02] magic_ninja_work, gives output but not related to the drive [03:02] neither dmesg [03:02] is not recognized at all [03:03] So the usb subsystem isn't seeing it at all. The most likely cause is that it isn't getting power / isn't getting enough power. [03:11] hello [03:11] anyone on 19.10? [03:11] i can't drag and drop stuff =( [03:11] i sware the drag and drop thing worked on 18 and earlier but could it be a virtual box problem? [03:23] anyone in ubuntu? [03:23] i knead help [03:24] comet23: so? [03:24] yea so liek the drag and drop dozen work in 19.10 is that normel? [03:25] comet23: no, it should work [03:25] hmm [03:25] i mean liek when u drag nd drop from downloads to desktop as example === MekelySpooky is now known as Mekely [03:57] got the printer working, eric pointed it out to me kind of earlier, but i wasn't extracting the tar file that was my driver in terminal and it didn't create a directory for the ppd files when I installed the install.sh file [03:57] i right clicked and extracted and that messed everything up [03:58] man that was painful [05:04] anyone know if there is a way to configure an hp scanner to scan to a linux machine from the scanner itself? [05:06] it looks like hp calls the feature "Scan to" [05:06] magic_ninja_work: what does that mean? [05:06] Most modern scanners have a list of computers on the local network you can send a scan to, and the computer is configured to receive and store it. [05:08] magic_ninja_work: that means your scanner got network connection? [05:08] Well yea [05:08] A good portion of modern printer/scanners are network printer scanners [05:08] mine is connected through usb. [05:09] I *can* connect mine through usb. It is just easier to plug it in somewhere and set it out of the way. [05:09] well, I'm sure it's possible. But you need to find out what protocol is used. [05:10] perhaps cifs and the receiving computer provides a share? [05:10] or do you need to install custom software on each computer? [05:10] I'm looking now. HP has a pretty good linux driver that just works. On debian I had to install it. On ubuntu it was already there and the cups daemon auto configures it. [05:11] But it also installs every single printer it finds, EVER, so I turned that off. [05:14] The funny thing is that the scanner detects my VMs and can scan to them. [05:27] maqic ninja work, install 'XSane' that little app will search for scanner [05:28] how do you install drag and drop to ubuntu 19.10? [05:28] like i want to drag downloads to desktop [05:28] but it doesn't work [05:39] comet23, install gnome shell extension --> https://extensions.gnome.org/extension/2087/desktop-icons-ng-ding/ will allow to drag and drop files on your desktop [05:40] thank you leden === JSmate1234 is now known as rymate1234 === MrSpooky is now known as MrC [06:14] magic_ninja_work: The sane driver for the HP scanners is for initiating the scan from the computer. The "scan to" function on the scanner usually drops files to an SMB share. [06:23] vlt, okay, that makes sense. So if the printer can see the share, then it will likely show up. [06:27] magic_ninja_work: That I don’t know. Maybe you have to manually provide the SMB config to the scanner. I always use the sane driver and refuse to even have a look at scanners that want it the other way round :D [06:30] the scanner works fine with sane, but when scanning multiple pages it would be nice to be able to do it all from the scanner [07:37] magic_ninja_work: with my scanner (usb connected), the buttons don't work. When I want to scan multiple pages, I use the "scanimage" program. The scanning speed is limited by the speed of the scanner tray. [07:38] or whatever that moving part is called [08:02] EriC^, moritzdietz: I ended up formatting the disk to ext4, and I'm trying again to copy the files on the 19.10 machine, however it starts fast and then the speed drops in the kB/s range and gets stuck there [08:03] I tried both the USB3 ports available, and also disconnected the usb mouse to make sure it doesn't interfere [08:16] hello! I am trying to find some easy open-source library software (clientside only) to use for the library in my school [08:16] (school I work at) are there any that exist in the default repositories? recomendations? [08:22] I dont manage to make any working udev rules on my system :S. Like this one she does not work. anybody knows why ? [08:22] SUBSYSTEM=="input", SUBSYSTEMS=="usb", ACTION=="add", RUN+="/bin/date > /home/max/test.txt" === Grldfrdom_ is now known as Grldfrdom === Grldfrdom is now known as Grldfrdom_ [08:52] haishu: your favourite web browser? [08:52] haishu: the library would then have a server side. [09:18] hi [09:18] anyone had problems connecting bluetooth headphones to ubuntu 18? [09:20] zamanf, I wish I could help with that issue. I am, at the moment, on Fedora 31. No bluetooth issues here. When I go back to Ubuntu I may be able to help. Apologies [09:22] coz_: if you cant help, please dont reply, leave the issue for other volunteers that can help [09:24] lotuspsychje, I understand what you are trying to say === im0nde_ is now known as im0nde1 [09:28] https://i.imgur.com/5UcKgaJ.png [09:28] Hi, has anyone an idea why it is not possible to use hdparm -y /dev/sdc in crontab but it works when use the command via normal shell [09:32] Netmage: you don't get a mail telling why not? [09:33] maybe give the full path to hdparm? /sbin/hdparm [09:36] This is what I see in syslog [09:36] Nov 1 10:00:01 ubuntu CRON[17356]: (root) CMD (hdparm -y /dev/sdc) [09:36] Nov 1 10:00:01 ubuntu CRON[17357]: (root) CMD (hdparm -y /dev/sdb) [09:37] Netmage: i was asking about mail not log [09:37] Netmage: can you run "mail" ? [09:37] my root user didn't get any mail [09:38] mail is not installed [09:50] Hi, [09:50] I am wandering why only after i did 'update-initramfs' the blacklist I have in /etc/modprobe.d took effect. I thought it has nothing to do with initramfs. Can someone explain that ot me? I am using 18.04. [09:52] Guest40456: perhaps the module was loaded inside the initramfs? [09:55] Wulf: Yes, I found the module inside initramfs but I wasn't aware the blacklisting in /etc/modprobe.d has an inpact on initramfs behaviour. Because it does ? [10:00] The complete path was the solution [10:00] 18.04, v5.3.0-21-lowlatency, on reboot lost access to embedded (USB) Bluetooth device. sysfs reports the device (hci0) but bluetoothctl nor hcitool find it [10:10] https://i.imgur.com/5UcKgaJ.png anyone knows how to fix this issue? it disconnects right after it gets visible (sony headphones) [10:11] Netmage: i was telling you so since the beginning [10:11] Netmage: and the reason is simply because cron doesn't get your users environment variables [10:14] zamanf: check in terminal> pactl list short | grep blue #you should see module-bluetooth-discover [10:15] Yes, this was the conformation that you are right [10:15] this was a implicit "thank you" [10:17] How can I set the correct environment variable for root in the crontab (-e) [10:19] i suppose you could export them from .xinitrc? [10:21] jeremy31, https://paste.ubuntu.com/p/mttjcvfhVC/ [10:22] zamanf: in bluetoothctl try> connect 70:26:05:AB:67:02 [10:23] Device 70:26:05:AB:67:02 not available [10:24] Netmage: I haven't read your actual problem but sometimes I solve cron/environment problems with a crontab line like this: * * ... /bin/bash --login the_actual_cronjob.sh [10:25] zamanf: was the headset still in pairing mode? [10:26] zamanf: you may have to do this in bluetoothctl> pair 70:26:05:AB:67:02 then> trust 70:26:05:AB:67:02 before connect [10:47] please how to set home page on chrome in ubuntu. i am lost :/ [10:48] vlt: Ok. I will try it [10:49] freelancerbob, https://support.google.com/chrome/answer/95314?hl=en === bootato is now known as fotato === yn_ is now known as yn [10:55] has anyone ready workaround for 2 monitor setup with different sizes? 4k and FHD monitors are nearly twice the size difference so moving mouse from one to another needs some aiming [10:56] is there a good workaround so mouse is smartly handled if the edge is detected? [10:56] so instead of getting stuck in a corner, it shows up on a second monitor edge and continues to move ? [11:15] At a loss here. An 18.04 desktop system has today repeatedly hung the GUI (nouveau) and eventually (after several minutes) been killed with no clues in any log, no processes consuming CPU/RAM, etc. I've done 5 poweroff restarts so far and it has happened every time. Any suggestions how to diagnose ? [11:24] TJ-, only thing I can think of is hardware; I'd use a 'live' system to see if issue occurs there (if it does there too, consider hardware tests), if it doesn't then maybe config/installed-18.04 somehow [11:25] TJ-: which nvidia graphics card? tried with nvidia binary drivers? [11:25] https://nouveau.freedesktop.org/wiki/FeatureMatrix/ [11:26] Hi, I'm having trouble to connect to a L2TP server. I'm getting `g_dbus_method_invocation_take_error: assertion 'error != NULL' failed` error. Here is the syslog https://paste.ubuntu.com/p/Rfk5CN2TqF/ Could someone please tell me how can I fix my configuration and connect to that server? [11:31] tarzeau: not touching nvidia drivers. Currently experimenting to see if the last 5.3 kernel update is responsible. Starting to look like it is [11:32] TJ-: how many screens does nouveau driver support at one card? 2? 3? 4? [11:33] (just being curious, since i've had so much trouble with them, including no support for CUDA) so no chance to use nouveau here [11:33] tarzeau: not entirely sure; I've got systems with 6 displays but all of those are 2 outputs-per-GPU instance [11:33] and that works with nouveau? [11:33] so their matrix must be heavily outdated [11:33] tarzeau: but from what I've seen in code and reports there's no problem handling all outputs the GPU presents [11:34] So far 4.15.0-68 is behaving... I'll start loading more applications and monitor it === frankndew is now known as deww === PowerTower_121 is now known as PowerTower_120 [12:09] I've got bionic on my home server and it seems to have a small delay when I ssh into it (compared to a previous distro). I'm thinking it might be the MOTD message displayed on log in? How can I disable this? [12:09] g1itch: how small is the delay? [12:10] about 2 seconds? [12:10] check for dns issues. [12:11] Why would it be DNS? It's a LAN SSH connection [12:11] (not saying it couldn't be) [12:11] because it's always dns. [12:12] I was having a similar issue with my FTP server within the LAN. It was DNS related. [12:13] I feel this is appropriate? https://www.cyberciti.biz/media/new/cms/2017/04/dns.jpg [12:13] So I'm running AdGuard on a local Pi as my DNS resolver. [12:13] Which hasn't changed from the last distro I had on this machine. === deadk is now known as e === TalkingMuffin is now known as Pici [12:56] gluon: it's MOTD, disable this: /etc/pam.d/sshd:33:session optional pam_motd.so motd=/run/motd.dynamic [12:56] oops [12:57] g1itch ^^^ [13:18] Since updating smbclient and related packages from 2:4.7.6+dfsg~ubuntu-0ubuntu2.11 to 2:4.7.6+dfsg~ubuntu-0ubuntu2.13, I can no longer access DFS shares. Is anyone else having this issue? [13:19] i see "CIFS VFS: ioctl error in smb2_get_dfs_refer rc=-19" in dmesg === mouseleums is now known as mouses [13:48] touch x y; ln -sf x z; ln -sf y z; will change symlinks and do the right thing: z-->y [13:49] mkdir -p x y; ln -sf x z; ln -sf y z; will not change symlinks, so ln -f does not work directories? isn't dir also a file under linux [13:49] a dir is a dir [13:49] a file is a file [13:50] how can I forcefully change symlink? do I have to remove the dir first or is there a flag like '-f' for directories [13:50] rr123, man ls says "List information about the FILEs (the current directory by default)." [13:50] it took me a long while to find out this bug :( [13:51] toolz, ^ [13:53] toolz, also see: https://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap03.html#tag_21_03_00_15 [13:53] ln -snf will do it, gosh [13:54] -n, --no-dereference treat LINK_NAME as a normal file if it is a symbolic link to a directory [14:01] rr123, if the target is an actual directory, you're not going to replace that with that command [14:01] you're aware of that, right? === im0nde1 is now known as im0nde [14:09] TomyWork: not really, I did "mkdir -p x y; ln -sf x z; ln -sf y z" and it worked for me, z-->y as expected instead of z-->x [14:09] i mean: ln -snf x z; ln -snf y z [14:11] rr123, yeah but you didn't mkdir z :) [14:13] i c, yeah I am not going to overwrite a real dir anyway, thanks for pointing this out. [14:28] good morning everyone [14:38] hi, random question. I'm having electricity cuts like crazy today so to keep up working I switch on my LTE that way I can still work on my laptop. The thing is, I cannot connect on my AWS/DO instances anymore via SSH? It hangs and ends up with connection closed. I have a work VPN on which I always have the same IP so this should not matter and my pub/priv keys are still there and ok. [14:39] shreds: That sounds like a network provider issue, not an Ubuntu issue [14:39] only thing ssh debug tells me weird is: Missing privilege separation directory: /run/sshd [14:40] my LTE network works fine, I'm here talking to you [14:40] my cable network is down because of electricity but that's not part of the question, only context [14:40] shreds: Some mobile carriers block encrypted traffic. You should call the network provider if all other things are working with out issue [14:40] by encrypted I'm talking VPNs and SSH [14:41] would I still be able to connect to my encrypted VPN if they'd do so? [14:41] VPN works fine [14:41] I can push to GitHub/GitLab via SSH fine [14:41] shreds: I'm not your mobile network provider, I have no idea [14:41] only connecting to my AWS instances on which the IP I'm using is whitelisted [14:41] I'll try to figure it out [14:41] thanks [14:41] shreds: Then that sounds like an AWS issue... if other things work. [14:41] process of elinimation there [14:42] shreds: are you sure your LTE provider isn't using CGNAT in which case the IP address your device gets locally may not be the one being seen by AWS/DO [14:42] servers are up and worked via SSH when I was on cable 20 mins ago, same for digital ocean [14:42] shreds: if you want to check hit one of those whatismyip sites [14:42] icanhazip give the right IP [14:42] so I guess not [14:43] but good point [14:43] shreds: can you connect to those endpoints using other protocols? [14:44] shreds: also, are you (trying to) connect to AWS/DO through your work VPN? [14:45] http/80 works [14:45] and using vpn to connect to aws worked 20 mins ago via my cable connection [14:46] only difference is what's behind the VPN now is my lte connection [14:46] shreds: http://ipchicken.com verify the public ip you're coming from [14:46] that's what boggling my mind, to the other end (AWS) it shouldn't matter if the private key and IP matches AWS' conf? [14:46] shreds: so locahost > LTE > VPN > work > AWs/DO port 22 = fail ? [14:47] shreds: or is it localhost > LTE > AWS/DO port 22 = fail ? [14:48] localhost > LTE > VPN > SSH PORT 22 for everything = fail [14:48] digital ocean, aws, my own dedicated servers, everything via ssh [14:48] I used AWS as example but testing anything else also fails [14:48] shreds: but *where* is that "VPN" terminating? at your "work", or at AWS/DO ? [14:48] but localhost > "CABLE" > VPN > SSH PORT 22 for everything = works [14:49] vpn is hosted on AWS [14:49] and works fine from LTE [14:49] shreds: I still really don't see how this is related to ubuntu at all. This sounds like a VPN/AWS/network provider problem [14:49] which makes it even weirder [14:49] This is related to SSH and related to my distro I used it on [14:50] shreds: so localhost > LTE > VPN > AWS connects. At that point you can access the AWS instance with protocols other than ssh/22, or not ? [14:50] shreds: have you confirmed the local routing table doesn't have entries that are pushing those ssh/22 connections out over LTE instead of inside the VPN tunnel? [14:51] I can always access AWS via all ports with or without VPN. Only SSH fails once I'm on LTE VPN or not. Sorry this is hard to explain I'll try to explain it better [14:51] localhost > LTE > VPN > SSH not working && localhost > LTE > SSH not working === ben_r_ is now known as ben_r [14:52] I did not check the local routing table [14:52] shreds: local firewall rules? are you using UFW ? [14:52] UFW is disabled on my workstation [14:53] how does it not work exactly? can you eg ssh with -v? [14:53] lte and a vpn seems likely to end in mtu mess [14:54] "connection closed" implies there was a "connection opened" [14:55] oops, I definitely didn't read through enough scrollback! [14:55] let me try with -v, I'll send the logs once it times out [14:59] holy cow sorry it works without VPN so only don't work "localhost > LTE > VPN > SSH" [15:02] shreds: if the packets are routed inside the VPN then they should get through unless the LTE provider is blocking it... what VPN software are you using? wireguard, openvpn, ipsec ? [15:04] openvpn [15:04] oh ssh -v is interesting [15:06] it hangs for a while, I'll pastebin the results as soon as I get it to timeout [15:08] shreds: can you do anything TCP-based through the VPN to the same target ? [15:08] web port (80) works I can curl it without issues [15:10] the VPN could be filtering port 22 upstream from you (unless you own the VPN and know it isn't doing that) [15:10] I own the VPN and it's as basic as it gets, straight openvpn but good point [15:12] @TJ- here are the logs https://hastebin.com/emutamivik.makefile [15:12] it hangs on debug1: expecting SSH2_MSG_KEX_ECDH_REPLY for a while and get connection closed [15:16] shreds: that is strange... connection established, Key Exchange started... I concur with tds who said earlier this may be MTU related since at that point the hosts will be sending larger packets to do the key exchange negotiation [15:18] I'm far from being an expert but this is an interesting lead. So IT connects it's just the handshaking that goes wrong. This boggles my mind since I'm far from understanding how and why that problem would be MTU related. I'll go read about that some more, interesting problem [15:19] same thing seems to happen with other servers, aws or not [15:19] shreds: see https://unix.stackexchange.com/questions/124946/ssh-session-through-openvpn-cuts-off-locks-up-after-a-few-lines and do some MTU/MSS testing and setting [15:20] shreds: what is the current openvpn MTU? "ip link show dev tun0" (assumin tun0 is the VPN) [15:20] lol we found the same MTU related posts :P that really seems to related damn you guys are better than me hehe [15:20] @TJ- 31: tun0: mtu 1500 qdisc fq_codel state UNKNOWN mode DEFAULT group default qlen 100 link/none [15:21] 1500 I guess [15:21] there are a lot of layers to this mess [15:21] shreds: 1500 is expected on a pure Ethernet link. when a VPN is involved tht should be less [15:21] the issue probably lies with the mtu on the lte interface, at a guess [15:21] shreds: just to try a hack fix, do "sudo ip link set dev tun0 mtu 1280" [15:22] tds: yes... shreds can you pastebin "ip link show" for us? [15:22] shreds: after the "...mtu 1280" command retry your SSH connection [15:25] ip link show: https://hastebin.com/yepuvekolo.cs [15:25] trying mtu hack [15:25] WTF [15:25] it works!!! [15:25] you are crazy geniuses [15:25] they all now work DAMN [15:25] shreds: ha, and that last pastebin doesn't show an LTE network device :) [15:26] I connect to LTE via my iphone tethering, sorry I forgot to mention that [15:26] shreds: you can configure openvpn to use a lower MTU so that it works whether it has LTE underneath or not [15:26] is there anywhere I could read a quick crash course why this resolves my issue? [15:26] shreds: oh... well ^^^^ [15:26] ohhhh good thing [15:26] so I can tweak openvpn config to specify MTU? [15:27] shreds: TCP packets that are marked "Don't Fragment" and are larger than the MTU ... what do routers do with them since they cannot forward them? [15:28] I guess they just can't do anything thus why it was timeout'ing? [15:28] oh this makes so much sense! [15:28] shreds: are you using Network Maanger to configure the openvpn client side? [15:28] yes and nmcli to up/down the connection [15:29] shreds: ok, so run nm-connection-editor select the VPN connection, on the VPN tab press the Advanced button, then in the new dialog's General tab set "Use custom tunnel Maximum Transmission Unit (MTU)" [15:30] on that note I have to run (literally) got 2 Huskies howling to go! [15:31] perfect [15:31] thanks SO MUCH [15:31] have a wonderful run with your huskies! [15:33] so happy I can work now :D :D :D thx tds as well! [15:36] Does ubuntu still allow mounting overlayfs in unprivileged user namespaces? I know there was a privilege escalation attack using those two features, but I can't tell if the attack was fixed or if overlayfs was just disabled in user namespaces? [15:51] Hello guys, Im looking for the lightest ubuntu distro to use in a raspberry 4 as a main computer, Im a front endeveloper so I commonly use vscode, chrome, firefox, a tons of terminal plugs like tmux and zsh, docker and node [15:52] i hope that usb bug is solved for rasppi 4.. [15:52]