[00:03] Hey how can I properly uninstall my nvidia drivers? They are not working, and are raising hell on my computer, and I want to opt for the drivers provided by nvidia. [00:03] it complains however that my kernel already has distro nvidia drivers, and wont install them. [00:04] go into the driver menu, switch back to nouveau, and reboot [00:04] that is, if you used the original repo drivers [00:07] Woodpecker: Find the .run file for the uninstall operation: sudo find / -name "NVIDIA-Linux-*" - you have to run the .run file again with --uninstall: sudo NVIDIA-Linux-x86_64-331.38.run --uninstall . For example. [00:08] hmmmm okay [00:08] wonder why they are not working [00:09] oerheks: it beats me. When selected on them, I had to boot to a commandline, and it was spitting out this constant message that made typing almost impossible. [00:09] Im also getting this error W: Possible missing firmware /lib/firmware/nvidia/gp102/sec2/image-1.bin for module nouveau === Wulf4 is now known as Wulf === urdh_ is now known as urdh [01:14] Im giving up on installing the manual nvidia drivers [01:15] it errors out because I am using Kernel 5.0 [01:17] Woodpecker: Why are you doing that ? [01:30] Hello! I'm having problems with my wifi, it's not recognizing the adapter. This happened before when I was using 19.04 and did this steps http://ubuntuhandbook.org/index.php/2019/04/nstall-rtl8723de-wifi-driver-ubuntu-19-04/ but now, after upgrading to 19.10 I'm having the same issue again. As far remember with dkms this shouldn't have happened [01:30] even after a kernel change. I'm running Ubuntu Studio. [01:32] My laptop uses the Realtek RTL8723DE [01:36] I guess my only question is, if I follow those steps again, will I have a problem? I'm asking because I already did it once. [01:38] Diebuntu: it might be weird if you have two git clones of that module sources [01:38] Diebuntu: doublecheck with dkms status before adding another copy [01:38] How do I do that? [01:39] run "dkms status" and see what the output .. if the driver is already listed, then maybe investigate. if it's not listed, then keep going [01:39] It is listed rtlwifi-new, 0.6, 5.0.0-32-lowlatency, x86_64: installed [01:40] lsmod? dmesg? [01:41] Should I paste the output on a pastebin or am I looking for something in particular? [01:42] for lsmod, your module; for dmesg, a pastebin might be better, it's hard to say what exactly might indicate a relevant error [01:42] Im on sitting_here(10.8.0.16) using vpn connection to the VPN server so i could connect via ssh to over_here(10.8.0.17)… from over_here i could access the webserver @over_there(192.168.8.8)... how to im on sitting_here to access all services provided @over_there via @over_here? Should the traffic conf be set on vpn server or on the client or even [01:42] on the over_here? [01:42] https://paste.ubuntu.com/p/XXDVJpB2vc/ - I must say right now I'm connected through usb tethering with my android. [01:45] Not entirely sure what I'm looking for with lsmod but I don't think I see my module, I do see a realtek though https://paste.ubuntu.com/p/6tR45xDdjv/ [01:51] [ 24.017004] Bluetooth: hci0: RTL: rtl: loading rtl_bt/rtl8723d_config.bin [01:51] [ 26.238297] Bluetooth: hci0: command 0xfc20 tx timeout [01:52] those lines seem to show that an attempt was made to load firmware for your bluetooth controller but the attempt timed out [01:53] I believe that's related too, bluetooth and wifi are the two issues I had to solve on 19.04. [01:53] hmm... but line 19 of lsmod gives me the feeling that the bluetooth may be working with a different module .. [01:55] i guess it probably needs an updated lfinger driver [01:55] ifinger? [01:55] larry finger [01:55] oh god. this again. [01:56] https://github.com/lwfinger/rtlwifi_new [01:57] tomreyn: oh nice. though I can't tell from the first two paragraphs if it'll help :) [01:58] me neither ;) but so far it seemed to help whenever the chipset was listed [01:58] and that's the case here [02:00] Thanks, but I'm still unsure of what I should do lol [02:01] Diebuntu: basically you'd follow the article you linked initally again [02:03] hi all [02:03] I see, is there a way the driver is installed in the previous kernel only or something like that? I don't know about this. I'm just guessing. [02:03] The previous kernel was 5.0.0-32-lowlatency and right now I'm Linux 5.3.0-19-lowlatency x86_64 [02:05] so step 3 of this guide gets the latest driver version from github.com and step 4 sets it up on your system, using dkms, allowing for this very driver version to be built against your updated kernel images. [02:06] however, this doesn't necessarily cater across releases since you will get a much newer kernel version (not just patches to a main kernel version, as you get within an ubuntu release). [02:06] I want to try the solution at the bottom of this page: https://www.dell.com/community/Inspiron/Linux-Inspiron-13-7380-wont-sleep/td-p/7283666 [02:07] Is it generally safe to set GRUB_CMDLINE_LINUX="mem_sleep_default=deep"? [02:07] Diebuntu: so you actually need to update the driver source code occasionally, such as on every new ubuntu release. [02:08] I see, yeah that makes sense. So I should follow from step 3, right? [02:08] k_sze: yes, undoing that if it doesn't work should be pretty easy, and if it dfoesn't work, you're probably only going to lose unsaved data when testing the suspend, which you likely already knew :) [02:08] Diebuntu: this is indeed not as it's supposed to me, you shouldn't have to work with github repositories. but unfortunately that is the current state of support for the chipset you bought there. [02:09] how can i install a package that is meant to run on 18.04 ? https://nvidia.github.io/nvidia-docker/ [02:09] Diebuntu: so in the future, you may want to shop more targetted, buying a piece of hardware where the manufacturer really cares for good linux support. [02:10] Yes, that's what I have in mind for my next purchase. I'm actually thinking of buying a laptop with ubuntu preinstall. Sadly when I bought this laptop I had no choice, I needed one fast lol. [02:11] ph88: only software in ubuntu is supported here (it could be in ubuntu, you could search for it), for other software please contact the developers through the support channels they provide. [02:12] mat88: because the ubuntu nvidia proprietary drivers are not working for me. [02:13] hmm, I have /dev/ubuntu-vg/swap_1 but it doesn't show up in the output of `mount`. Is that normal? [02:14] I's a 1 GB Swap (version 1), according to gnome-disks. [02:14] s/I's/It's/ [02:14] When trying to follow step 3 I get fatal: destination path 'rtlwifi_new' already exists and is not an empty directory. [02:16] Diebuntu: delete the existing directory recursively [02:17] before you dkms install, be sure to remove the old one [02:17] Diebuntu: ^ [02:18] Oh I'm not sure I know how to do that [02:18] dkms remove rtlwifi_new [02:19] actually dkms remove rtlwifi-new [02:19] maybe even sudo dkms remove rtlwifi-new/0.6; sudo dkms remove rtlwifi-new [02:20] Output is: Error! Invalid number of parameters passed.Usage: remove / --all or: remove / -k [02:20] I assume I should remove the 0.6 [02:21] yes, i think so [02:21] Still same error [02:21] so what have you run exactly? [02:21] !pastebin [02:21] 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. [02:22] https://paste.ubuntu.com/p/39Nktk2vm5/ [02:22] k_sze: https://bugzilla.kernel.org/show_bug.cgi?id=199689 suggetss that the 5.3 kernel should do better on suspend to ram [02:22] bugzilla.kernel.org bug 199689 in Hibernation/Suspend "s2idle does not work in Dell XPS 9370" [Normal,Closed: code_fix] [02:22] Diebuntu: ah sorry we'Re just missing --all [02:23] No 5.3 kernel for me since I'm still on 18.04. [02:23] I wonder if I should upgrade to 19.10. [02:23] k_sze: it's there via poposed in hwe-edge [02:23] !proposed [02:23] The packages in Ubuntu are divided into several sections. More information at https://help.ubuntu.com/community/Repositories - See https://wiki.ubuntu.com/RecommendedSources for the recommended way to set up your repositories. [02:24] So... sudo dkms remove rtlwifi-new/0.6 --all [02:24] Diebuntu: so thiat's sudo dkms remove rtlwifi-new/0.6 --all [02:24] yes [02:25] Yes, it's going through [02:25] tomreyn, odd. The word "proposed" doesn't appear on either of those pages. [02:26] Looks like that means pre-release. [02:26] It's done, I should follow from step 3 now, right? [02:27] k_sze: https://wiki.ubuntu.com/Testing/EnableProposed [02:27] k_sze: yes it means pre-release [02:29] Diebuntu: yes. but before you do step five let me check whether it should still be 0-6 [02:29] * 0.6 [02:31] How do I tell where my swap is coming from? `free` says I have 999420 MB of swap, but `mount` doesn't show any swap partition mounted. [02:32] Diebuntu: yes it's still 0.6 [02:32] k_sze: cat /proc/swaps [02:33] `/dev/dm-2 partition 999420 0 -2` [02:33] Awesome, thank you! I'll follow the steps now [02:35] I'm still getting the same error, is it the folder I need to remove? https://paste.ubuntu.com/p/gF8x42WZ3c/ [02:35] k_sze: lsblk /dev/dm-2 [02:35] Diebuntu: delete the existing directory recursively [02:35] so rm -r ./rtlwifi_new/ [02:36] tomreyn, ah, ok. so it *is* ubuntu-vg--swap_1: ubuntu--vg-swap_1 253:2 0 976M 0 lvm [SWAP] [02:36] k_sze: then this is ... good ... i guess? === nt0_ is now known as nt0 [02:37] Why doesn't it show up in `mount` at all though? Conventionally, non LVM2 swap would show up in the output of `mount`. [02:37] It's a bit confusing. :/ [02:38] i don't think so. i think swap never shows up in "mount" [02:40] Diebuntu: does it seem to be working out? [02:40] Yes, so far so good [02:40] Thank you [02:40] I'm on step 6 [02:42] Diebuntu: when you're done, check if you already have a file /etc/modprobe.d/rtl8723de.conf [02:43] i.e. cat /etc/modprobe.d/rtl8723de.conf [02:45] Yes, I do have a .conf file with antenna 2 selected. [02:45] And it already connected to the wifi [02:46] So wifi is working, but I still don't have bluetooth adapter. [02:47] try a reboot [02:48] maybe even a full power-off reboot -- I wonder if the timeout while loading firmware may or may not persist across 'simple' reboots but might be affected by a full power off [02:48] k_sze: actually you no longer need -proposed for 5.3 in ubuntu 18.04 since today, it's now the proper hwe-edge image [02:49] I would like the print manager to open and iconise on the taskbar when I print. Instead of having to browse to the printer settings and view the print manager there. Is it possible, and if so could someone point me in that direction please [02:49] ? [02:50] Diebuntu: i'd comment out the line in the /etc/modprobe.d/*.conf file about this module, since it may now work out of the box for you. only re-add (uncomment) it later if needed. [02:50] I'm sorry I got disconnected, I'm not sure I was able to read all [02:50] And now I can't login with diebuntu lol [02:51] Diebuntu93: now you can [02:53] Woodpecker: Did you tried install the nvidia drivers from ppa ? [02:53] mat88: mmmmm I think I might have [02:53] mat88: the only reason I say think, is because I was getting a signature error when running them [02:54] which I thought might have meant a ppa signature [02:54] yes [02:54] http://ppa.launchpad.net/graphics-drivers/ppa/ubuntu is in my sources [02:55] I'm back, where can I check the logs to read the previous messages? [02:57] Diebuntu: I cut and pasted a few lines to https://pastebin.com/j6P6dh30 [02:57] !irclogs | Diebuntu: here, but it's updated with a delay of ~ half an hour [02:57] Diebuntu: here, but it's updated with a delay of ~ half an hour: Official channel logs can be found at https://irclogs.ubuntu.com/ . LoCo channels are now logged there too. Meetingology logs at https://ubottu.com/meetingology/logs/ === zbenjamin_ is now known as zbenjamin [02:57] Whoops, tomreyn ... did not mean to step on toes. [02:57] wodemoncafe: that is weird. Well I say that because from my experience installing the binary downloaded from the nvidia site is the worst experience [02:57] mmm Im going to try disabling the source, and installing nvidia driver 390. [02:58] okay, I'll try the reboot now [02:58] Ascavasaion: you didnt step on mine there, no worries [03:02] tomreyn, eh, `apt search hwe-edge` doesn't show the kernel image, only headers and tools. [03:02] (that's after already having `apt update`'ed) [03:06] k_sze: linux-generic-hwe-18.04-edge [03:06] I'm back, wifi is working, bluetooth is still not found. [03:07] tomreyn, ah, ok. Thanks [03:07] Diebuntu: can you share your logs? journalctl -b | nc termbin.com 9999 [03:07] Diebuntu: and did you comment out ant_sel? [03:07] journalctl -b | nc termbin.com 9999 [03:08] No, I'm not sure what you meant by that. [03:08] Diebuntu: when you're done, check if you already have a file /etc/modprobe.d/rtl8723de.conf [03:08] i.e. cat /etc/modprobe.d/rtl8723de.conf [03:08] Yes, I do have a .conf file with antenna 2 selected. [03:08] Diebuntu: i'd comment out the line in the /etc/modprobe.d/*.conf file about this module, since it may now work out of the box for you. only re-add (uncomment) it later if needed. [03:09] Sorry, didn't paste the link https://termbin.com/q09z [03:09] tomreyn, does edge eventually get "promoted" to non-edge? [03:11] looks like it does. https://wiki.ubuntu.com/Kernel/RollingLTSEnablementStack#hwe-16.04-edge [03:11] Diebuntu: actually, look like i misread what you wrote about what ant_sel is set to, so just ignore this part [03:12] k_sze: yes it does [03:14] Diebuntu: is bluetooth locked via rfkill? [03:17] Is this what you mean https://paste.ubuntu.com/p/NNqqW3JSD7/ [03:17] oh the timeout is still there: kernel: Bluetooth: hci0: command 0xfc20 tx timeout [03:17] yes thats what i meant [03:18] but it's not inactive [03:18] do you see it listed in settings -> bluetooth? [03:18] i.e. are you able to switch it on in the top right corner there [03:19] Yes, I can turn it on and off, but when I try to setup a new device I get a pop up saying "no adapters found" [03:22] i see. maybe another driver is needed for the bluetooth [03:27] Okay, so i should look for RTL8723DE bluetooth driver, right? [03:28] Diebuntu: i'm not sure really. *maybe*. what does htis output? ls -l /lib/firmware/rtl_bt/rtl8723d* [03:28] https://pastebin.com/dYi9dwjs [03:31] Diebuntu: ah there was a recent update to the firmware [03:31] https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/log/rtl_bt/rtl8723d_fw.bin [03:32] I see, so I need to update that one too? [03:33] I'm having trouble changing the background of the login screen (Lock Screen?). Under settings -> background, I set one of the images to be background and lock screen, but the lockscreen continues to be that violet color with the ubuntu logo. [03:33] Diebuntu: sudo -i then cd /lib/firmware/rtl_bt; mv rtl8723d_fw.bin rtl8723d_fw.bin.orig; wget https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/rtl_bt/rtl8723d_fw.bin ; exit [03:35] Done, should I reboot? [03:36] Diebuntu: reloading the module would be enough, but rebooting should also work if this fixed it [03:37] i assume the module is called rtl8723d or rtl_bt [03:38] you'd sudo modprobe -rv rtl8723d rtl_bt to remove both modules (if those exist), then sudo modprobe -v rtl8723d [03:39] Nope, not found modprobe not found [03:39] modprobe: FATAL: Module rtl8723d not found. [03:39] modprobe: FATAL: Module rtl_bt not found. [03:39] well then maybe rebooting is faster ;) [03:47] I'm back, still no adapters found. [03:49] Diebuntu: hmm, can you run: sha256sum /lib/firmware/rtl_bt/rtl8723d* /lib/firmware/rtlwifi/rtl8723defw.bin [03:50] https://paste.ubuntu.com/p/m5CV9q4p8w/ [03:52] Diebuntu:hmm yes those are the latest [03:52] Diebuntu: lets check the log again: journalctl -b | nc termbin.com 9999 [03:53] https://termbin.com/zbgd [03:59] Diebuntu: so the bluetooth issue remains this - https://paste.ubuntu.com/p/3RMP5GDsDF/ - i'm afraid i can't seem to find a solution to this problem right now, and also need to get some sleep [04:01] Yes, I understand. You've been very helpful [04:01] Thank you a lot! [04:01] you're welcome. [04:42] Hello Everybody [04:43] What is the best way to install ubuntu along with Windows nowadays ? [04:43] Is Wubi still the go to option ? [04:52] Got a wee bit of a problem. I've a server up and running, and it constantly drops down in performance, so ping just skyrockets for no reason. [04:52] But! I found that it correlates to the clock speed. The ping skyrockets when Linux has the CPU go down to idle speeds [04:53] How do I stop that from happening, so Ubuntu doesn't downclock the cores? [05:02] Fifty, try: sudo cpupower frequency-set --governor performance [05:03] also try schedutil instead of performance [05:03] Says command not found, unfortunately. [05:03] Tried to install that one earlier, and the package can't be found when doing sudo apt install cpupower [05:03] I think I found the ticket with cpufreq-set, tho: http://manpages.ubuntu.com/manpages/bionic/man1/cpufreq-set.1.html [05:03] you'd need linux-tools [05:04] and probably linux-tools-$(uname -r) [05:06] Hmm. Alright, I tried that. But the terminal says that it's a virtual package, and offers a bunch of oddly named...Things. Screenshot: http://manpages.ubuntu.com/manpages/bionic/man1/cpufreq-set.1.html [05:07] try the cpupower command again, should be there now [05:07] Nope, says command not found. [05:08] However, I did manage to change the governer with cpufreq-set. But don't see any difference, and it didn't print anything to the terminal [05:08] did you do all the cpus? [05:08] that's why cpufreq-set sucks. cpupower will do them all for you [05:09] Err, probably not. I just typed in: sudo cpufreq-set -g performance [05:09] how many cpus on that? [05:09] There's 4 cores, 8 threads. [05:09] But didn't pass anything like that to the command [05:10] this will do them all: for c in {0..7}; do sudo cpu-freq -c $c -g performance; done [05:10] That's a bash, aye? Lessee if I can get that to work [05:10] it is [05:11] oops, cpufreq-set actually [05:11] not cpu-freq [05:12] Right, saved that to a .sh, ran it, don't see any diff the now via lscpu [05:12] I'll join the server, see if it still drops [05:12] Thanks for the help so far [05:13] you can use turbostat to monitor cpu throttling/boost closely [05:14] or sudo atop [05:15] if you really want wide-open-throttle, you can pass a minimum frequency in the cpufreq-set loop, with --min [05:15] usually that's a bit overkill but it can be used to rule out boost pr governor as the cause [05:15] Aye I was thinking of doing that. [05:16] Setting it to governor still causes the frequency to fall, unfortunately. [05:16] linux's "performance" governor is not comparable to windows. windows' sets min to the base freq, linux performance governor just makes it more apt to max it out but still throttles down when it is near idle [05:18] it's probably not the governor causing a problem if I had to guess though. but not a bad idea to rule that out at least by measurement [05:19] ping is very very easy on the cpu. next to nothing [05:19] Aye just ruling it out like. I've done all the obvious stuff with wine commands and such, now I got to do this stuff. [05:19] you can probably saturate your NIC with your cpu at min freq [05:19] What happens is, the server software suddenly has it's ping skyrocket, due to it's main loop (displayed as Server FPS) suddenly taking longer [05:20] I noticed that it directly correlated to the clockspeed, and others have had the same issue with Wine applications [05:20] ah, game server? ya set min to the max frequency then [05:20] games hate governors [05:22] I'd run cpufreq-info, look at the frequency range it reports "within X and Y" and make that cpufreq-set --min Y [05:23] Hmm. The clockspeed still drops below what I set --min to [05:23] I'll give that a shot [05:24] Aye it says between 1.6G0 GHz - 3.90 GHz [05:24] And is using the intel_pstate driver [05:25] then --min 3900M should do it [05:26] It should, but it still drops. Bizarre. Even did it manually in the terminal. Changed the .sh batch file to: for c in {0..7}; do sudo cpufreq-set -c $c --min 3400; [05:26] Literally just that in the .sh [05:26] how are you watching the frequency? [05:27] if using turbostat, look at Bzy_MHz column [05:27] With watch -nl [05:27] watch what [05:28] I'd use sudo turbostat and look at Bzy_MHz column [05:28] Ah, sorry. Let me get the command [05:28] Gotta copy paste as it's got that funky line symbol [05:29] Here we are. The command I punt into the terminal is: watch -n1 "lscpu | grep 'CPU MHz' | awk '{print $1}'" [05:29] I'll give turbostat a run [05:31] Agh, it's part of those linux utils [05:32] why is that a problem though? [05:32] Couldn't figure out which one to install and how [05:33] I suggested already: sudo apt install linux-tools-$(uname -r) [05:33] that will automagically use the one for your kernel [05:33] Oh I didn't see that [05:33] My bad [05:36] Aye according to Byz_MHz they're around 3500-3600MHz [05:36] I'll try now and see if that fixed it. Thanks again. [05:36] good luck :) [05:38] Damn, that didn't fix the issue. [05:38] what does this say: sensors [05:38] (to check cpu temperature) [05:39] The highest it's seen for core temp is 58 [05:39] Package temp 60 [05:39] that's fine [05:40] A big possible fault is that I have to run it through WINE. [05:40] Native linux game servers run fine. Just this one for Conan Exiles has shoddy performance [05:40] can you do this and try to reproduce: watch -n 0.5 sensors [05:41] under load cpu temp can skyrocket. checking at idle is almost meaningless [05:41] or better yet, have that running and in another terminal: stress -c $(nproc) [05:41] Oh what I do, is I make the PuTTY terminal always on top, then hop into the game with this PC [05:41] will peg every cpu [05:41] Gotcha [05:42] Unfortunately, "watch -n 0.5 sensors" doesn't work. [05:42] "Sensors not found" [05:43] probably need: sudo apt install lm-sensors [05:44] Aye there we go [05:44] Temperatures are fine, doesn't go past 60 with the server running [05:44] Lets see the stress test [05:45] the mass of a cpu die is so low, it can shoot up 30C in a split second under load. I've been fooled by idle temps before [05:46] Don't have the stress command either. The server is under light load, about 10% on each thread. [05:46] had a machine that looked like it never went over 60C that actually shot up to 90+ under pegged load [05:47] sudo apt install stress [05:47] Oh my. Restarted the server, and one of the cores reached 70C when loading everything [05:48] 70 is not too bad. it will start halting at around > 80 though [05:48] depends on the exact cpu [05:48] It's ivybridge, so it can take those temps [05:53] another thing to try is: sudo perf top [05:54] actually, sudo perf top -z [05:54] reproduce the performance issue and see what shoots up to the top. it might tell you what's doing it [05:55] Good idea, I'll do that the now [06:02] 28% perl perl_scalarvoid23% perl perl_yyparse18% perl Perl_slab_alloc20% perl Perl_leave_scope8% kernl __schedule [06:02] There were other spikes that I couldn't copy down in time [06:02] Ew it broke formatting [06:02] https://pastebin.com/bvp0tDhu [06:03] This as well: 25% libc.2.29.so [06:04] As well as: 32% perl Perl_Op_Free [06:05] Perl seems to be the culprit. There's loads of Perl shared objects, but a couple of them suddenly reach high CPU usage for a few seconds [06:13] strange, what would be doing a lot of perl in that scenario? [06:15] No idea, to be quite honest. It doesn't even use perl, it's built on Unreal Engine 4. [06:15] perhaps it is watching for filesystem changes and getting sprayed with inotify? [06:15] is the program in your home directory? [06:15] It's in /home/steam/servers/conanexiles [06:16] Now I just restarted the dedicated server, no software running. [06:16] Yet perl is still using a bunch of the CPU. [06:17] Same with libc: https://i.imgur.com/GIFS4Yf.png [06:17] Very strange. [06:18] I'm guessing the perl is probably something running in the background unrelated to your workload [06:19] bigger guess: and your workload is causing tons of file changes in your home directory, causing the perl to work hard chasing after all the changes [06:19] That's the thing. The Conan Exiles server isn't even running. [06:20] And it's getting the same Perl CPU usage. [06:20] ah [06:20] Is there a way to see which service or process is running that perl stuff? [06:21] run htop, press t, look for perl in the list [06:22] press < and choose sort by PERCENT_CPU [06:23] ah, tree view defeats the sort [06:23] in non-tree view do you see perl near the top when sorting by percent cpu? [06:23] Gi'es a moment, trying to figure out what I'm exactly looking at here :P [06:24] list of running processes [06:24] Oh aye, i gathered that much [06:24] Just overwhelmin [06:24] if you apply the sort by cpu with <, the most relevant ones will be at the top [06:27] Odd, it's not showing anything sucking up the CPU in htop, when the perl usage spikes happen. [06:27] Screenshot: https://i.imgur.com/Yeiu9wa.png [06:29] looks like it is x2go using perl though [06:29] Oooh. [06:30] heavy screen changes cause spikes maybe? [06:30] I'm not even connected to it at the moment. [06:30] I wonder...Can I kill the process somehow? [06:30] Or, err, service. The x2go thing. [06:31] sudo service x2goserver stop [06:33] Now CPU usage shown in htop is 0% for all the cores, the kernel now being in the high percentages on that perf top window [06:33] If x2go was the problem, I'm gonna scream. [06:38] Nope, still getting that performance drop. [06:38] are you actually using mysql for something? [06:39] if not: sudo service mysql stop [06:39] That's for another server program. Not using it right now though, couldn't get it to create the MySQL databases yet. [06:39] Good idea [06:39] any partitions near full? check: df [06:39] Nope, they've plenty of space. [06:39] ext perf goes through the floor when near full [06:40] I've 2x1TB, and only used around 50GB total [06:41] ssd write performance can go awful when near full too [06:41] ssds? [06:41] Nope, using HDDs. That did come to mind a few hours ago tho. [06:42] Contacted Hetnzer, see if they'll let me buy an SSD. [06:42] Got a screenshot when the performance cratered again: https://i.imgur.com/Bb3oa2B.png [06:43] This time, it's the Kernel that's gone red in top [06:44] wow, that's an awful lot of timer interrupts [06:45] I wonder if that's due to one of these launch commands for the server program [06:45] I'll try removing it [06:50] I'm not getting as many interrupts, but the performance still craters. perf-3404.map appears to be a culprit. https://i.imgur.com/d2XSJp4.png [06:50] Err, perf-4304.map rather [06:51] Kernal as well, with the symbol import_iovec https://i.imgur.com/Zzy0nl9.png [06:54] what does this say: uname -r [06:55] 5.0.0-32-generic [07:01] how about: ps -p 4304 === drcd_ is now known as drcd [07:12] PID - 4304, TTY - pts/3, TIME - 00:13:04, CMD - ConanSandboxServer [07:12] So half of it is due to that .map thing [07:31] Fifty, that's the actual workload though right? probably not much you can do at host level about that, you'd have to look into performance issues with that program itself under wine [07:32] Aye, sounds about right. It's just sad that it runs better when hosting on my own desktop, while also playing the game. Couldn't find a thing about using it with wine [07:43] Thanks for the help so far. Gonna head off to bed the now. === Leion1 is now known as Leion [08:36] I have a fresh install of Ubuntu 18.04 server with Freeradius installed from the Ubuntu repos. On boot the Freeradius service doesn't start. [08:38] what does systemctl status have to say? [08:38] or journalctl [08:39] systemctl freeradius status [08:39] Unknown operation freeradius. [08:40] otherway around, it should be systemctl status freeradius [08:40] zetheroo, status comes before the service namm,e [08:40] that order would work if you invoke "service" [08:41] https://pastebin.ubuntu.com/p/syQB32ZjCg/ [08:41] the service starts fine if I do 'service freeradius start' [08:41] but doesn't start after reboot [08:44] what about journalctl -u freeradius.service [08:44] should tell a little more info [08:46] https://pastebin.ubuntu.com/p/V5RhVG7wCk/ [08:46] looks like the output of /var/log/freeradius/radius.log [08:47] hmm it seems to start fine, then be forced to restart by 'radius freeradius[19142]: rlm_sql (sql): Removing connection pool' until it meets the amount of restarts allowed by the service file and fails [08:48] you'll have to follow that up with freeradius, because I'm not familiar with it and i cant see anything that could be going on wrong with the service file configuration [08:49]