[00:02] jeremy31: I'm nearly 1.5hrs into disabling bt_coex and I still haven't had a drop. So while I can't be conclusive yet, I do have some targeted exploration to do over the next couple day [00:02] s/day/days [00:03] HALP MEH [00:06] ld_library_path= will break csgo from launching [00:07] i dont know if i should put /usr/lib32 or /usr/lib/libSDL-2.0.so.0 [00:07] /usr/lib32/libSDL-2.0.so.0* === VileGent is now known as kk4ewt [01:14] new to ubuntu. i'm trying to print to an epson artisan printer, but the 1430 driver is not listed in cups. having a hard time figuring out how to find out which package to install....any ideas or good search terms? [01:23] I have device conneted to a USB port, lsusb shows it as there, and as Bus 001 Device 049: ID 0525:a4a2 Netchip Technology, Inc. Linux-USB Ethernet/RNDIS Gadget, ip a shows it as connected and up - there is an sshd server on that device, how do I connect to it [01:25] ssh using local host [01:28] tonyt: if that was meant for me, I tried ssh themachinename.local and get "can't resolve hostname" _ know I have the name rightsice if I let the conneted evice connect to the internet, THEN I can connect to it that way [01:29] if localhost doesnt work try the assigned internal ip [01:29] *the=then [01:29] tonyt: by localhost do you mean the .local that I am adding? [01:30] to me localhost is 127.0.0.1 [01:30] ya [01:31] tonyt: the device is not even pingable - it is 192.168.1.206 - or so IT thinks [01:32] tonyt: so connectng by ip does not work [01:36] jpmh, well, if connecting by ip doesn't work, there's no point it trying to figure out why .local name resolution isn't working. is this a multi-homed host (more than one network interface connected?) [01:38] rfm: do you mean the device that I am having toruble connecting to? if so,yes, he has a wlan and the dhcp file places that on 192.168.1.205 and usb0 on 192.168.1.206 [01:38] except clearly it does not [01:40] the device clearly has the two interfaces since the wifi connection works find and my laptop even sees the non-wifi connection as a gadget [01:43] jpmh: does the ip's the device gives itself match the same type of ip's on your local network? (192.168.1.x) [01:44] leftyfb: yes - everything is 192.168.1.x - the machne that I am on is .198, and the wifi on that device is 205 and the usb is supposed to be 206 - in fact, that machine can ping to itself on 206 [01:44] jpmh: that is why you can't connect [01:45] jpmh: you can't have 2 separate networks sharing the same subnet [01:45] jpmh: reconfigure the device to handle a different ip range. Like 192.168.0.x or 10.0.0.x === zbenjamin_ is now known as zbenjamin [01:45] leftyfb: ty - that's what I was missing [01:46] leftyfb: being called, so can't test not, but I'm sure that is iit - ty [02:11] is there a tool to get a filesystem diff periodically? Would using a journaling filesystem help in that regard? I'm hoping to just get a text output of all changed files since the last run, and then hopefully run it every 24hr [02:12] jayjo: to what end? [02:13] I would like to email it to myself to monitor a server [02:14] monitor it for what reason though... [02:14] anyway [02:14] there are packages just for this purpose, like AIDE [02:14] if you wanted to write your own simplistic one [02:15] depending on the number of files and FS, it might be easy enough to just 'find /some/path | sort | tee files."$(date +%Y%m%d)" [02:16] jayjo: if you've got zfs set up, it's pretty easy http://paste.ubuntu.com/p/Jn3yBXR85Q/ [02:16] and then diff -s <(ls -1rt files.* | tail -2 | head -1) <(ls -1rt files.* | tail -1) [02:18] why would i would want to build cairo with gles2 support only and no opengl? [02:19] most likely you wouldn't [02:19] that sounds like a phone thing [02:19] but it's for "embedded" systems, yeah [02:19] hmm [02:19] so it's a lesser implementation [02:19] but on arch i saw opengl is disabled [02:19] https://en.wikipedia.org/wiki/OpenGL_ES [02:19] is opengl idsbaled on ubuntu for cairo? [02:20] Also, sometimes when installing software with make, I'd like to be able to compare entire fs diffs to understand what went where. I've been using docker instead and have been getting started with lxc to keep things isolated, and considered that safer. Is there a better way to approach that? [02:20] IDS is definitely on my mind, I'm looking into AIDE more [02:20] shoober420: maybe a wayland thing [02:21] shoober420: that is [02:21] a way to use wayland without pulling in X [02:21] or more specifically, to use weston [02:21] hmm, thats a possibility [02:21] but, y'know, wayland [02:21] wouldn't surprise me if Arch is farther ahead on wayland adoption than Ubuntu [02:22] * matsaman shrugs [02:22] this is a sway/wayland/radeon setup [02:22] but i would prefer opengl over opengles [02:22] im using steam through xwayland since steam still depends on gtk2 [02:24] if you have X stuff already, then probably using opengl instead will not cause any additional deps [02:24] wayland can do native opengl, but not when the program uses gtk2 [02:25] you could probably use i3 and not be such an early wayland adopter, too [02:25] as boring as that'd be [02:25] i use bspwm on x11 [02:26] i still x11 with my nvidia machine [02:26] radeon and sway/radeon [02:26] i actually bought two cheap dell precision 3620 towers for cheap and put a geforce 1660 super in one and radeon 5500 xt in aother [02:27] shoober420: what version of ubuntu are you running? [02:27] ok fine i use gentoo [02:27] shoober420: please stay on topic then. This is an Ubuntu support channel [02:28] shoober420: feel free to chat in #ubuntu-offtopic [02:29] ok === mIk3_09 is now known as mIk3_08 [06:29] * enyc meows [06:30] found the cat [06:33] mew [06:42] hi, my sound card seems to get put to sleep when not used for a while in ubuntu. any way to prevent this? [06:43] it seems that it is woken up when being put back to use and it screeches during the wakeup [06:48] what sorta card? [06:48] what sorta sound card? [06:50] generic builtin one on a msi motherboard [06:50] snd_hda_codec_realtek hdaudioC1D0? [06:55] I tried this, seems to work https://www.reddit.com/r/archlinux/comments/7vz8k6/since_recent_update_linux_puts_sound_to_sleep_on/dtylrm9?utm_source=share&utm_medium=web2x&context=3 [06:55] I don't know how to preserve it between updates though [06:56] **between reboots [07:05] nshire: this is a laptop? [07:05] no, desktop [07:06] nshire: do you have tlp installed? dpkg -l | grep -i tlp [07:07] it appears not [07:09] nshire: might try this first: https://unix.stackexchange.com/questions/565886/how-to-disable-the-power-saving-for-snd-hda-codec-realtek [07:09] and then look at https://forum.manjaro.org/t/unable-to-disable-snd-hda-intel-power-save-through-modprobe/31568/5 [07:10] if you wanted to just take your echo command and use that [07:10] look at https://askubuntu.com/questions/919054/how-do-i-run-a-single-command-at-startup-using-systemd [07:10] or you could simply use cron [07:19] matsaman: thank you, I commented load-module module-suspend-on-idle. rebooting to check if it worked. [07:22] seems to have worked === PowerTower_121 is now known as PowerTower_120 [07:30] anyone around? shot in the dark question here: [07:31] yeah [07:33] have 2 machines (ubuntu laptop on wifi) and a desktop on a local net. every day i need to ssh into my laptop, but always get a bad response "Destination not found"...so, what it comes down to (after some troubleshooting and not being able to reproduce the issue until the next day) i just ping the laptop and wait for a good response, then ssh [07:34] some details: the laptop has a net connection (interface is up and connected), i am not logging in to X11/gnome on the laptop (thought that it mattered but didn't seem to make a difference) [07:35] found this, which seems like it could be related, but i am on 20.04: https://askubuntu.com/questions/1095566/unable-to-ping-18-04-desktop-unless-this-machine-first-pings-the-client-when-us [07:39] i turn the laptop off every night, then when i start it in the morning the same issue happens. if i reboot the laptop now and ping/ssh it all works fine [07:53] Hi, how can i disable KDE autostart after the system boot ? [07:53] I'd like to try another WM [08:14] XATRIX: It should be SDDM display manager I think? Maybe check your /etc/sddm.conf for autologin and comment it out. [08:15] overclucker: thanks [08:17] XATRIX: no problem :) [09:13] Hi. It seems that the legacy images for pxe have been removed, but software such as foreman relies on them being available. Now, you get redirected to some fancy HTML page if you go there. What to do? [09:15] afaiu if you want to keep using old ubuntu releases, the pxe images remain where they always were, is that not correct? [09:15] if you want new ubuntu releases, you're supposed to switch to -live-server .iso images, they support PXE booting [09:15] https://ubuntu.com/server/docs/install/netboot-amd64 [09:16] Yeah, I know, but they are _very_ bloated and big, what happened to efficiently small netboot setups? [09:17] sadness happened [09:17] :'| [09:18] the best bit is how cloud-init also fetches the netboot image two additional times because it misinterprets the url parameter on the kernel command line [09:19] So that's 3GB? [09:19] superschnell: mgedmin: Having set this up recently, I ended up going with a local http server and mirroring the contents locally. Goes faster :) [09:19] 3*1.2 GB, but there's a workaround for the cloud-init thing [09:19] I can't find the thread on https://discourse.ubuntu.com [09:20] I can't believe this is happening :( [09:20] Oh well [09:20] superschnell: Also, to save you some time, if you do uefi+secure boot PXE boot, make sure you grab the focal-updates grub shim, not the focal one. Has a bug on some ethernet adapters. [09:21] https://discourse.ubuntu.com/t/please-test-autoinstalls-for-20-04/15250/292, but there was also a separate thread on that subtopic [09:21] Thanks [09:21] ah, here: https://discourse.ubuntu.com/t/ubuntu-20-04-installer-downloads-iso-more-than-once/21697/3 [09:22] and I linked to an irrelevant message in the middle of the thread because I don't know how to use discourse [09:22] https://discourse.ubuntu.com/t/ubuntu-20-04-installer-downloads-iso-more-than-once/21697 [09:37] I'm trying to use a nfc reader/writer with python, but it appears it's being used by another service in ubuntu [09:37] pcscd which I think is used for another smart card elsewhere [09:37] how can I disable it taking over this usb device? [09:38] why dont you talk to pcscd instead ? [09:39] oh hrm, I killed it and it's not effecting my usage at the moment [09:39] I probably was testing with it earlier. [09:39] ogra, it's more of a general thing [09:40] ah, I was experimenting with firefox / chrome smartcard usage [09:42] cluelessperson, well, do you want to write your own pcscd in python ? that'd be quite some work (pcscd interfaces with all the low level drivers) .... i'd simply use python3-pyscard to talk to the daemon to not having to bother with low level drivers [09:43] ogra, I'm working on writing nfc to a specific chip, I only installed that for smartcard usage testing [09:44] well, your choice 🙂 [09:53] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1923670 yay, the fix for this is in the hwe-5.4 and hwe-5.4-prep branches for bionic now. now when's the next kernel release? [09:53] Launchpad bug 1923670 in linux-oem-5.10 (Ubuntu) "CIFS DFS entries not accessible with 5.4.0-71.74-generic" [Undecided,Confirmed] [10:55] gnask@gnaskBuntu:~/Hämtningar$ sqlcmd -S localhost -U SA -P 'passwordhere' [10:55] Sqlcmd: Error: Microsoft ODBC Driver 17 for SQL Server : Can't open lib '/opt/microsoft/msodbcsql17/lib64/libmsodbcsql-17.7.so.2.1' : file not found. [10:55] gnask@gnaskBuntu:~/Hämtningar$ cd /opt/microsoft/msodbcsql17/lib64/ [10:55] gnask@gnaskBuntu:/opt/microsoft/msodbcsql17/lib64$ ls [10:55] libmsodbcsql-17.7.so.2.1 [10:55] Can anyone help me with this? The file is clearly there but i still get an error, Ubuntu 20.04 === Scotty_Trees is now known as Scotty_Trees|Zzz [11:07] gnask: permissions issue or apparmor issue looks like [11:07] also, what does tat /opt/microsoft/msodbcsql17/lib64/libmsodbcsql-17.7.so.2.1 say? [11:08] correction: what does stat /opt/microsoft/msodbcsql17/lib64/libmsodbcsql-17.7.so.2.1 say? [11:27] how can i use gnome keyring in remote ssh sessions? [11:27] superschnell, Fil: /opt/microsoft/msodbcsql17/lib64/libmsodbcsql-17.7.so.2.1 [11:27] Storlek: 2063728 Block: 4032 IO-block: 4096 normal fil [11:27] Enhet: 805h/2053d Inode: 1330464 Länkar:1 [11:27] Åtkomst: (0755/-rwxr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root) [11:27] Åtkomst: 2021-04-29 20:41:28.727465548 +0200 [11:27] Modifiering: 2021-03-04 20:23:56.000000000 +0100 [11:27] Ändring: 2021-04-29 20:40:30.623162698 +0200 [11:27] Född: - [11:28] i got it to work through x2go, but i cannot get it to work in SSH [11:31] drag0nius: ssh -A remotehost [11:31] so what release will include gnome40 [11:31] but only for the part of gnome-keyring that acts as an ssh-agent [11:31] .... actually now that I think about it, does gnome-keyring still act as an ssh-agent? wasn't that part disabled a while ago? [11:31] i need it to work with git credentials over ssh [11:32] reading remote matchine's keyring [11:32] I generally use ssh keys as git credentials, for which ssh -A works great [11:33] for other kinds of credentials, hmm, if you use ssh -X to forward the X11 session, would that forward dbus too? maybe not [11:39] (you _really_ should warn about security issue when recommending ssh -A) [11:40] (this option expose your private key to privileged users on "rebound" host) [11:46] Hi folks [11:46] hey [11:51] -A doesn't expose the key itself, but lets root on the remote host use your private keys while your session is active [11:51] and yeah, I should've mentioned the security implications [11:56] greetings, how can I list only the physical drives in server? no partitions and no logical vloumes. Just physical drives attached mounted and not mounted. [12:02] isomari: what are you trying to accomplish? [12:03] you can look at 'ls /dev/disk/by-id/ | grep -v part' for example [12:05] or 'lsblk -a -d' [12:13] sorry if this is a bit off topic but is there a way to see all the ubuntu irc bot commands? [12:13] !brain | pizzaiolo [12:13] pizzaiolo: Hi! I'm #ubuntu's favorite infobot. You can search my brain at https://ubottu.com/factoids.cgi | General info and channels at https://wiki.ubuntu.com/IRC/Bots | Make a clone of me, see !botclone [12:14] ty! [12:38] Hi, can you advice ? I'd like to make a system for my kiosk. Is it possible to say kubuntu not to load, any displaymanager? But only od startx with openbox/fluxbox ? [12:39] I suppose it to run to "init 3" and then do some kind of a script with trigger 'startx openbox' or something thislike [12:50] I want to insert the date into a file: I do it like this: date "+%Y-%m-%d %T" > /tmp/my-faketime.rc [12:51] but I need to add a "@" at the beginning of the file: "@2021-04-30 12:46:58" [12:51] and Idea how to solve this in a simple cmd? [12:51] date "@+%Y-%m-%d %T" > /tmp/my-faketime.rc doesnt work [12:54] I'm trying to follow this https://ubuntu.com/tutorials/install-and-configure-wordpress#4-configure-database [12:54] But I get ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'IDENTIFIED BY ""' at line 4 [12:55] what line are you trying to run [12:55] what part of the process [12:55] mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP,ALTER [12:55] -> ON wordpress.* [12:55] -> TO wordpress@localhost [12:55] -> IDENTIFIED BY ''; [12:55] Query OK, 1 row affected (0,00 sec) [12:56] I made identified by into [12:56] mysql> GRANT SELECT,INSERT,UPDATE,DELETE,CREATE,DROP,ALTER [12:56] -> ON wordpress.* [12:56] -> TO wordpress@localhost [12:56] -> IDENTIFIED BY ''; [12:56] Query OK, 1 row affected (0,00 sec) [12:56] > IDENTIFIED BY ""; [12:56] I mean [12:59] You can also do - create user 'new_user'@'localhost' identified by 'new_password'; [12:59] Then - grant all on my_db.* TO 'new_user'@'localhost'; [13:06] thank you robertparkerx [13:11] !paste | katareeena [13:11] katareeena: 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. [13:11] sorry I double pasted by mistake [13:13] katareeena, usually anything over 3 lines is too much, that used to be the rule iirc [13:17] zen_coder: echo -n "@" > /tmp/fake; date "+%Y" >> /tmp/fake [13:24] zen_coder: you could also use sed to insert @ as first char, sed 's/^/@/' /tmp/fake [14:01] Hi guys. On my new AMD laptop (amdgpu) I'm experiencing what some is calling "false hotplug events" on HDMI. When my external display goes to sleep, it wakes up after a second or two. I have found many posts on this issue from a couple of years ago, but nothing recent. Does anybody know a good way around it? [14:02] tmus: Try turning off port auto detection on the monitor itself. [14:04] rjwiii, thanks, i already tried that. I'm not sure my monitor allows complete disable but with DDC/CI disabled, it still happens. [14:04] Worked for me ... I have 3 monitors. [14:05] hi all! [14:06] in ubuntu 20.04, is there a way to get Nvidia 460.73.01 via apt? [14:07] rapha: always use ubuntu-drivers list to see wich drivers available for your cards chipset [14:08] rjwiii, alright cool. I'll have to figure out a way to properly do this on my Lenovo G34w-10 monitor then. Soulds like the way to go... Could the cable make a difference? And do you know if this is _only_ a problem on HDMI or USB-C as well? [14:09] lotuspsychje: thank you! it seems there's two available of that version, one with "-server" at the end ... should i use that because i'm headless? [14:10] rapha: headless on plain server or a desktop? [14:10] lotuspsychje: it's a libvirt VM running on a 20-core Xeon without a physical screen attached. [14:12] rapha: if you dont need the graphical violence might need the -server driver [14:13] heheheh yeah no need for graphical violence, this is just for CUDA :P [14:13] rapha: some users tend also to download drivers from the ubuntu graphics ppa, for latest versions [14:13] !nvidia [14:13] For nvidia and matrox graphics cards, see https://help.ubuntu.com/community/VideoDriverHowto . For AMD/ATI graphics cards, see « /msg ubottu ati » and « /msg ubottu fglrxmissing » For the latest Nvidia drivers see https://launchpad.net/~graphics-drivers/+archive/ubuntu/ppa [14:21] ah, latest features isn't necessary. i need latest production, which is 460. [14:21] allrighty go for it rapha [14:21] xubuntu 17.10 - uptime was over 100 days and after reboot it's no longer resolving internal domain names using our dns server. systemd-resolve --status lists the our dns servers correctly. if we do nslookup internal.domain.blah.com it is unable to resolve but if we do nslookup internal.domain.blah.com 10.1.2.3 which is the dns server listed in the status, it resolves it fine. resolv.conf is a linked file and it's set to 127.0.0.53 [14:21] uggestions or ideas appreciated [14:22] groklinux: 17.10 is eol [14:22] i understand i just need to fix this though [14:23] groklinux: then test it with "host .... 127.0.0.53" [14:24] yeah that came back with multiple entries [14:24] groklinux: systemd-resolve --status|grep -i "dns servers" [14:24] groklinux: then your applications mysteriuously are using an old or broken '/etc/resolv.conf' [14:25] groklinux: are you using 'nscd' too? [14:25] so kill it and do the dpjg-reconfigure? [14:25] not to my knowledge [14:25] groklinux: systemd-resolve --status|grep -i "dns servers" [14:25] groklinux: also look at "leftyb" command line. [14:26] I personally try to use 'unbound' as the local resolving server. [14:26] the grep listed the correct internal server [14:26] groklinux: only the 1? [14:27] https://utcc.utoronto.ca/~cks/space/blog/linux/LocalDNSCacheMaybe [14:27] yes only one [14:27] Hi i have to enable secureboot and i am wondering how to make dkms to sign modules. i read https://wiki.ubuntu.com/UEFI/SecureBoot/DKMS and woud love to follow method 1. as of now i have SB disabled and have 4 dkms packages installed. how can i rebuild/reinstall them without SB being already enabled? [14:27] i have 2 internals and 8.8.8.8 and 8.8.4.4. [14:27] Walex: lets not recommend non-standard configurations to fix an issue [14:27] groklinux: ok, then is your issue [14:27] groklinux: remove the externals [14:29] groklinux: client resolvers should all be configured to respond in the same way since they all get queried randomly. 8.8.8.8 and 8.8.4.4 will not resolve any of your local DNS hosts, but those nameservers will be used when trying to query local hosts and fail [14:30] groklinux: if your local (internal) resolver is set to forward requests to some external nameserver like 8.8.8.8 and 8.8.4.4 then you don't need to set those on the clients [14:30] groklinux: especially if you need to resolve local hosts [14:34] sorry for the delay...removed the externals using the network-configuration gui (which was fine before) and status is still listing all 4. verified when i reopened network-configuration tool that they are not listed...i restarted the resolved service (naively) but it's still listing all 4 [14:34] groklinux: where did you remove the externals from? [14:35] ipv4 tab of network-configuration gui in xubuntu [14:35] groklinux: is the machine getting it's ip address via dhcp? [14:36] no, it's set [14:37] i am thinking rm the resolve.conf and do the dpkg-reconfigure resolveconf might be next? [14:37] groklinux: do not do that [14:38] roger [14:38] groklinux: if you want this fixed, just follow along and do not make any changes without being directed [14:38] will do [14:38] groklinux: reboot [14:39] ok will have to give me a few as i have to notify the floor (they're prepped but not ready) [14:40] groklinux: why is such a critical server running a desktop environment? [14:40] groklinux: also, over 3 years EOL [14:41] if you have worked for a very large corp, you will already know how budgeting and false promises of IT works === ufk is now known as ufk_ [14:42] groklinux: 6 months to a year, sure. 3 years going on 4, no. Not acceptable [14:42] we never in a million years expected this to be doing this now in 2021...never...we even have the real big iron for it sitting right beside it...it's a tragedy tbh [14:42] groklinux: we all know, and I know places with a couple of racks of Windows 2003 servers... :-) [14:42] lol exactly...there's no excuse but it's reality i have to live in [14:43] BTW XUbuntu is nice for servers: It is convenient to have a lightweight GUI on the console. In part because of that I have often used *laptops* as critical servers (laptops have builtin UPS, builtin console, very robust, easy to relocate in a hurry). [14:43] billion upon billions of dollars and heck, i have to ask irc friends to get help...you would think i'd have a pile of consultants at the least tripping over themselves [14:43] groklinux: once/if this is resolved, the next step is getting it upgraded to an LTS release. One that is supported for 5 years for fre and 10 years with ESM. The version you have there was only supported for 9 months. Not something that should have been put on a mission critical machine with no plan on keeping it up to date [14:44] yeah we have the plans and like i said, the iron is sitting there...this is literally running on and old dell desktop...and i told you how much money goes through this you'd fall over [14:44] Walex: there are zero reasons to install a GUI on a mission critical server unless there is some mission critical application that requires a GUI [14:45] regardless, back on topic. [14:45] ok rebooting [14:45] * Walex uses LTS in his personal laptop, as GUI upgrades often don't amount to much [14:47] * Walex uses HWE newer kernels with LTS though, as that is fairly transparent [14:48] ok it's done rebooting [14:50] groklinux: ok, check again: systemd-resolve --status|grep -i "dns servers" [14:50] Can anyone help me diagnose crossftp not launching in Ubuntu. Installed the *.deb successfully, clicked to launch the application, pops up for a few seconds and then closes. [14:50] WE'RE BACK BABY!!! [14:50] oops...i did an nslookup real quick and it worked without including the dns server [14:51] skeeter: launch it from command line to see if there are any error messages. [14:51] skeeter: crossftp is not part of the officially supported packages. You'll need to seek support from the crossftp project [14:51] groklinux: make sure systemd-resolved is only listing your internal nameservers, otherwise you'll run into this problem again [14:52] good point [14:52] skeeter: anyhow there are several alternatives to CrossFTP [14:52] lfetyfb, any officially support packages you recommend? ftp clients. [14:52] just the 2 now leftyb [14:52] skeeter: I recently tried 'jftp' [14:53] skeeter: I don't recommend ftp at all, but if you must, I have used ncftp in thr past [14:53] api's are able to pull from the aws stuff again...good deal. [14:54] just so that i can get my head right...we did have the 2 externals in there prior to the reboot without issue...it appears taking them out resolved it but i'm unsure as to why [14:54] hi :) in focal, is there a way to have the openssh-client read a config file from a pipe? with "ssh -F <(./guess_config.sh) web.local date" it says "Can't open user config file /dev/fd/63: No such file or directory" [14:54] groklinux: entropy [14:54] skeeter: for FTP only there are very many, command line, 'curses', GUI, and many frameworks have builtin support for 'ftp://' [14:54] LOL i'll take that answer [14:54] Sven_vB: why would you want to do such a thing? [14:55] groklinux: in the goold old times DNS servers were used in the order they were listed in, unless there was an option to rotate them. I suspect that 'systemd-resolved' now does rotation. [14:55] leftyb and the rest, please enjoy a the rest of your friday...once again, irc friends come through where all the money in the world cannot [14:55] groklinux: as in, client resolvers hit the different configured nameservers randomly. There is no order or weight. You need to make sure all configured nameservers will response to all queries in the exact same way with the results you expect [14:56] leftyfb, to read the appropriate private key path from a database by hostname, rather than writing all of them into my config, or need a directory of full of hostname-named symlinks. [14:56] or hostname-named config files to include [14:57] Sven_vB: what is the use case? I don't quite understand what you are doing. Openssh uses it's ssh_config or ~/.ssh/config for it [14:57] also optionally more options like proxy and timeout. [14:57] it's "database" with no way to use an external db [14:58] Sven_vB: if it is just to specify the private key path why not use '-i keypath' option? [14:58] leftyfb, querying the DB is my bash script's job. I could of course put lots of options on the command line, it just feels more elegant to use the config file syntax. [14:59] Walex, ^ [14:59] Sven_vB: ssh -i $(find-path-of-key-for ...) .... [14:59] Walex: that's not what they are asking [15:00] Sven_vB: free free to create a shell script that does all the lookups and builds you ssh command. Pipes are going to work [15:00] aren't* [15:01] Leftyfb, do you know if JFTP supports PRET capable ? [15:01] skeeter: I have no idea. Feel free to look at the man page or project page [15:02] skeeter: note that IIRC GVFS supports several URL schemes... [15:02] skeeter: for FTP my favorite client is command line 'lftp' [15:03] skeeter: feel free to /join #ubuntu-offtopic for software recommendations [15:03] Thank you guys. [15:46] what issue does 21.04 have with booting? === EriC^ is now known as EriC^^ [15:46] when upgraded from 20.04? [15:47] !21.04 | donofrio [15:47] donofrio: Ubuntu 21.04 (Hirsute Hippo) is the 34th release of Ubuntu and is the current regular release. Download at https://www.ubuntu.com/download - Release notes: https://wiki.ubuntu.com/HirsuteHippo/ReleaseNotes [15:47] check the releasenotes for the known bugs donofrio [15:49] lotuspsychje, I would have but this is where I seen the issue from https://lists.ubuntu.com/archives/ubuntu-devel/2021-April/041448.html short answer has the shim been rolled into all upgrades now or? (context that email was from the 24th) [15:50] donofrio: the mailing list post you pointed to refer to ablocking bug, with url. when you access this url, you should see its state on the top of the page [15:51] see if you can work it out [15:51] tomreyn, it renders fine for me [15:51] donofrio: do you have an issue with 21.04 booting? [15:52] tomreyn, https://imgur.com/GUj1ZnV.png [15:52] leftyfb, yes [15:52] donofrio: ok, what issue are you having and what have you tried to fix it? [15:52] leftyfb, I mean I don't have an issue just asking before I upgrade my device.... [15:52] donofrio: please detail any error message [15:53] donofrio: ok, then you do not have any issues as of yet [15:53] leftyfb, no because I'm on 20.04 [15:53] donofrio: feel free to ask for support here when you come across an issue. Good luck [15:53] hence my question.... [15:53] leftyfb, now looking for a guess if you do not know if shim is in place then just say so... [15:53] now = not [16:05] donofrio: let me try to explain again: https://lists.ubuntu.com/archives/ubuntu-devel/2021-April/041448.html has two web links on the bottom. one of them is about the bug that's holding back release upgrades to 21.04 (as the text explains). [16:05] donofrio: i'm suggesting to access the bug report, and to examine its status. you can see that on top of a bug report. [16:06] "hirsute" is the release codename for Ubuntu 21.04. [16:07] https://www.reddit.com/r/pcmasterrace/comments/n1x1xe/google_play_store_should_a_qr_code_for_every_app/ [16:07] "Hirsutism is excessive body hair on parts of the body where hair is normally absent or minimal." from wikipedia. Hmm... [16:07] so... this release might be ... a hairy one...? [16:08] not here please cbreak_ [16:10] tomreyn, so you could have just said "I don't know" or "no progress on this EFI issue" instead of sounding smug - just sayin. [16:10] my device is EFI so I guess I'll hold off even though this was released as 0f 4/4/2021 [16:11] donofrio: i was trying to enable you to understand the state of a bug report on your own next time around. please don't just assume people are meaning to do you bad. [16:12] it's not that it is just this - there are 1k users here probably wondering perhaps the same thing so that is why I asked here should I have asked in the +1 dev channel instead (guessing not cause they already moved on to 21.10 guessing) [16:14] donofrio: it's certainly fine to ask your ubuntu support questions here. further discussing the topic of what should be discussed here should move to #ubuntu-ops, after you read the links in the channel /topic [16:14] wonder if upgrading to 21.04 is easier when systemd-boot is used, or zfs-boot-menu, instead of grub [16:38] hey, anyone aware of a tool that is able to monitor `ps` and kill a process matching a name when it goes above a certain number in memory usage? but only the process name that's causing it as there can be duplicates with the same name [16:38] I could prob write a script but want to avoid any edge cases if there is a pre-built solution already [16:39] Dev0n: shouldn't the proper solution be to fix the application that has a memory leak? [16:40] leftyfb, ofc, but the fix is still being explored as I don't know what's causing it but if I leave this running it ends up consuming the whole server and that ends up crashing [16:40] crashing, how? [16:40] Dev0n: I'm not aware of such a tool. Maybe write a wrapper for said application that monitors and kills as needed [16:41] it's best of the worst solution at this point, it's a job worker and anything killed will get re-added to the job queue so nothing is really lost [16:41] tomreyn, because it ends up leaking memory until there is no memory left [16:41] then the whole server just grinds to a halt [16:41] leftyfb, ok cool just checking [16:42] Dev0n: because it has swap and that's not configured well, so the OOM killer never kicks in, or because that's not configured well either? [16:43] tomreyn, I have configured it with limits to actually stop this from happening, however, the service I'm using is not respecting these limits, there are some edge cases that causes a task to escape out of those config params [16:43] it's a python celery worker, for context [16:43] and celery is notoriously bad for memory leaks [16:44] Dev0n: pretty simple; put the offending processes into a mmory control group (cgroup) [16:44] TJ-, ooh interesting, never worked with cgroups but something I can look into [16:45] Dev0n: here's a worked example https://www.monetdb.org/blog/limit_memory_usage_on_linux_with_cgroups [16:46] nice! thanks TJ-, this could be a very temporary solution while I try and debug the bigger issue at hand [16:47] Dev0n: if you're using either --system or --user systemd units then see "man systemd.resource-control" and "MemoryMax=" and related options [16:48] hah even better, I am actually using systemd to start the celery processes === docdevel2 is now known as docdevel [16:50] Do "snaps" update when you do "sudo apt upgrade" ? [16:50] no [16:50] How do you update them? [16:50] explodes: snaps update on their own [16:50] @_@ [16:50] explodes: to run it manually, you can run sudo snap refresh === docdevel is now known as docdevel2 [16:51] they do on update-manager GUI after apt updates [16:52] ok thanks all [17:01] TJ-, not seeing anything in the docs but I assume that MemoryMax= set on a service will kill "all processes" that gets started by that service right? [17:01] For example, in my case, the service starts celery which starts a number of workers, it's the workers that could be stuck leaking [17:02] Dev0n: the control group usually is set to apply to a group of tasks/processes, so if they're children of the service's master PID they should be contained. [17:03] Dev0n: https://www.freedesktop.org/software/systemd/man/systemd.resource-control.html "If memory usage cannot be contained under the limit, out-of-memory killer is invoked inside the unit" that should apply to any children process [17:04] TJ-, leftyfb ok the workers end up getting their on PID, there could be 10 workers and if one of those workers go above then looks like they would all get killed [17:05] Dev0n: looks like you would use MemoryHigh= in preference to MemoryMax= [17:06] Ahh TJ-, reading that option now, that looks like any workers that would be within the limit are unaffected and only the one that's going crazy is chopped [17:06] cool, guess that will work better in this instance [17:08] Dev0n: there's kernel docs on those at https://www.kernel.org/doc/html/latest/admin-guide/cgroup-v2.html#memory-interface-files [17:08] interestingly enough, someone is running to the same issue I'm having and looking at these options in their systemd https://serverfault.com/questions/1021898/memory-limits-in-systemd [17:08] Does anyone know if Ubuntu 21.04 will switch between internal GPU (Intel) and eGPU (nVidia)? I don't mean whilst the laptop is powered on. (e.g. power off, plug in eGPU, switch on, have wayland+nvidia) [17:09] TJ- ahh ok, these are for cgroups but if I'm doing this through systemd, those are abstracted away and systemd will take care of it right? [17:11] Dev0n: correct; the systemd options create/write to those though, so there is more detail in the kernel docs (or more context sometimes) [17:11]