[00:43] my seagates are like that, never caused issues [00:43] as far as I understand, modern drives are so dense they are correcting all the time. [00:45] as long as you dont have any pending or reallocated sectors it is prob fine === Tahvok_ is now known as Tahvok [01:03] trippeh: thanks for the data point [01:03] trippeh: the output on those seagate is weird and very diff than my HGST or any other disk I've seen really, tyhere's now raw data table [01:04] so I can't see the reallocated sectors for example [01:04] the link above suggested looking at "Elements in grown defect list: 0" [01:05] and several others did the same [01:05] these are seagate constellations btqw === Spydar007 is now known as Guest92796 === cpaelzer_ is now known as cpaelzer [06:18] While I'm not up to speed yet I wanted to mention that today is again Bug Squashing Day https://wiki.ubuntu.com/ServerTeam/BugSquashingDay [06:18] chair likely is just me for now [06:18] rbasak: are you around already? === Spydar007 is now known as Guest77180 [06:52] Good morning [06:59] hi lordievader, good morning to you as well === JanC is now known as Guest5365 === JanC_ is now known as JanC [07:12] hiya, is there a way to keep ubuntu from changing /home/$USER/.bash_history ownership back to root on apt upgrade / reboot, and purging the contents? [07:12] running ubuntu 16.04.2 xenial [07:13] i've manually reassigned ownership of that file to the user a couple of times, but after a reboot it seems to go back to root and have been blanked [07:14] Deeps: I have never once seen that. [07:14] Deeps: what have you installed that didn't come from the archive? [07:16] Deeps: have you done any funny configuration of ~/.bashrc or ~/.bash_profile? any funny pam modules? any funny /etc/profile or /etc/profile.d/ games? [07:16] best i can tell, nothing particularly exciting on there at all [07:16] http://paste.ubuntu.com/24458832/ [07:17] looks pretty normal [07:20] Deeps: you could install auditd and add a file watch rule on the file. once auditd is installed, edit /etc/audit/audit.rules and add a line -w /home/deeps/.bash_history -p wa [07:20] Deeps: then /var/log/audit/audit.log would contain information on the process that modified the file [07:21] http://paste.ubuntu.com/24458844/ like that? [07:22] looks good; I think now systemctl restart auditd.service [07:22] timestamping in the log file isn't very friendly, heh [07:23] no it is not :) [07:23] looks like it's working, i manually changed it back to ownership of me and got this in the logs http://paste.ubuntu.com/24458852/ [07:24] will see what resets it next time, thanks for the help [07:24] great, good idea to test right away [07:25] aureport -f looks like a friendlier output view. neat. [07:25] realised it's not a reboot that's doing it, as the box has been up 100 days and the bash history only got reset in the last week [07:25] oh that is nice, cheers [07:36] Hey cpaelzer, how are you doing? [07:48] lordievader: good enough to complain as I always do :-) [08:05] cpaelzer: Good, good :) [08:08] hey guys how can i get more info about this process: tcp 0 0 127.0.0.1:5005 0.0.0.0:* LISTEN 11500/6 [08:10] funabash1: 11500 should be the pid [08:11] funabash1: so look around at things like [08:11] ps axlf | less (search for 11500 in there) [08:11] ls /proc/11500/* - ususally comm, exe and stat* are interesting there to start with === haasn is now known as hanna [08:19] cpaelzer: check this please, https://pastebin.com/ePPGe532 [09:07] funabash1: back now, reading [09:09] funabash1: so the process is your sshd it seems [09:09] not the common port for ssh but well [09:09] funabash1: OTOH you might recheck if you still want to look for 11500, in case the pid just got reused by sshd [09:10] funabash1: the first you posted was netstat output right? [09:47] cpaelzer: yes, but the electrician's here so I may disappear suddenly :-/ === Tuor_ is now known as Tuor [10:20] cpaelzer: i did kill it [10:32] this is so strange,m preseeding servers using a 64bits only server cannot find some packages [10:32] but they are there in 64 bits [10:54] is it possible to remove dash and only use bash? [10:55] possible yes, sensible no [10:55] why not just use bash where you want, [10:58] k [12:27] cpaelzer: add bug 1683237 to your list of why you should have core dev please :) [12:27] bug 1683237 in krb5 (Ubuntu Zesty) "krb5-user: kinit fails for OTP user when using kdc discovery via DNS" [Undecided,Triaged] https://launchpad.net/bugs/1683237 [12:28] hehe, yeah correct [12:30] xnox: are you Mr. systemd now? [12:30] cpaelzer, probably.... [12:30] enough commitment :-) [12:31] xnox: fyi I have come to this while triaging server bugs (and duped something onto it) but wanted to make you aware of bug 1624317 [12:31] bug 1624317 in systemd (Ubuntu) "systemd-resolved breaks VPN with split-horizon DNS" [High,Confirmed] https://launchpad.net/bugs/1624317 [12:32] reading the comments there it seems that might get some extra hot-ness now that zesty is releases with systemd-resolved [12:32] cpaelzer, my networking know-how is low; thus most of the resolved/networkd bugs are beyond me =/ i do my best, but some help in that area would be good. [12:33] in like trianging and telling what things do. I can write code and send it upstream, but e.g. i have never knowngly ran split-horizon DNS or how that supposed to work correctly. [12:35] me neither, I just happened to follow the "dns leak argument" in the other bug until I realized they were the same [12:46] TafThorne: thanks for starting to verify the logrotate issue on Xenial [12:46] TafThorne: do you think you can do the other releases as well over the next few days? [12:46] cpaelzer: No problem. I'll need to fire up some fresh VMs after finding install images to do the others. [12:47] TafThorne: no pressure, just wanted to know if that task is with you [12:47] TafThorne: thank you in advance [12:49] cpaelzer: I should be able to install a Trusty and Yakkety Vm... although I might also have a Trusty PC somewhere here that I could test on easily enough. That would be faster. [12:50] TafThorne: have you tried lxd instead of vms? [12:50] Never before. I could try that [12:50] it's even faster [12:51] TafThorne: what's your base distro? [12:51] where you work? [12:52] TafThorne: http://pastebin.ubuntu.com/24459990/ [12:53] then import your ssh key, and you can ssh into it [12:59] ahasenack: 16.04 [12:59] ahasenack: Basingstoke, UK [12:59] TafThorne: should work just fine [13:00] ahasenack:cool thanks. That would be much faster. [13:01] ahasenack: Can I go down as well as up? So `lxc launch ubuntu:trusty` or somilar? [13:01] TafThorne: yes, even precise [13:01] or debian [13:01] or many others [13:06] TafThorne: if this is the first time you are installing lxd, the only gotcha is that you have to add yourself to the lxd group before using the commands, which means a logout/login sequence or some other trick [13:13] ahasenack: thank you for the warning. I have a terminal installing the client. I'll add myself to the group now. [13:15] (i'll have to remember the syntax for that before I add a user named TafT to a a group named useradd but I can work that out with GOogle in a few seconds) [13:19] I use gpasswd -a lxd [13:19] I have gone with `usermod -a -G lxd ` [13:21] Should be the same result. I'll not try and fancy jumping though hoops. I will just `sudo login` as me again in a terminal. That should be enough to confuse me tomrro wwhen I look at the console agian. [13:22] `groups` suggests I am added. Time to follow he paste. [13:25] ahasenack: That was quite quick to get started. http://pastebin.ubuntu.com/24460153/ Now I guess everthing else is like any other headless terminal. I go set-up to use -propose, install the package and run the test case? [13:32] TafThorne: yeah [13:32] it took 27s because it had to download the image, I had it cached already [13:32] TafThorne: try "lxc image list" and "lxc list" [13:33] TafThorne: since you didn't run "lxd init" before launching the container, I'm not sure how your networking is setup, lxc list should tell if your container got an ip or not [13:54] ahasenack: does not look like it has an IP [13:54] TafThorne: I'd suggest to tear it down then and then run sudo lxd init and follow through the setup wizard that it runs [13:55] ahasenack: will do. I'll get back to it when I am at a good point to logout and login agian to sort all that out too. [13:55] ok === drab_ is now known as drab [17:45] urm [17:45] I've been running qemu in console on my desktop for testing and it works fine [17:46] I moved the same script to a remote server and when I try to start it I get "Could not initialize SDL(No available video device) - exiting [17:46] " [17:46] however I was using qemu on my desktop from terminal, with -curses [17:46] so not opening the GUI [17:46] but you can't use -curses or -nographic with -daemonize (which seems to imply it) [17:47] so I'm a bit lost... I can't use the paramters to tell qemu to not start a graphical env if I demonize it, which makes no sense to me [17:47] any clue? [17:48] I also tried -vga none and still get the same error [18:04] drab: what's the command line you are using? [18:04] drab: you can test locally on your desktop by unsetting DISPLAY probably, that would replicate the remote case [18:05] drab: also try -display none [18:05] or -display vnc and the vnc parameters (see manpage), these don't require DISPLAY (the shell var, for X access) and work with -daemonize [18:06] ahasenack: /usr/bin/qemu-system-x86_64 -enable-kvm -name test -machine type=ubuntu,accel=kvm,usb=off -cpu host -smp 1 -m 2G -netdev bridge,br=lxdbr0,id=qemubr -device virtio-net-pci,mac=52:54:00:11:01:18,netdev=qemubr,id=eth0 -drive file=test/rootfs.raw,format=raw,id=root,if=none -device virtio-scsi-pci,id=scsi -device scsi-hd,drive=root -boot dn -monitor unix:test/mon.sock,server,nowait -pidfile [18:06] test/instance.pid [18:06] then with -daemonize at the end or -curses [18:06] right now I'm going through an install with -curses on the remote server for testing [18:07] so that works fine [18:07] ahasenack: I'll try -display none, thanks, vnc seems a good workaround too [18:08] drab: do you need the console to drive things interactively? [18:09] no, it's all preseeded, ultimately I don't want any console [18:09] ok, then -display none should work I think [18:09] trying, thanks [18:09] -nographic is another one [18:13] ahasenack: it worked, thanks a lot [18:13] \o/ [18:13] anothe strange thing I noticed testing [18:13] I have -boot dn [18:13] but it always seemst o boot from network [18:13] even tho the disk install is valid and -boot d will boot just fine [18:14] and with -curses I stil see a complaint about not being able to boot cdrom, even tho I never ask it to boot from cdrom [18:14] isn't "d" the cdrom? [18:14] oh, lol, I thought d was disk, n network [18:14] my bad [18:15] c cdrom [18:15] should have doubel checked [18:15] yeah, c is "disk c", from windows fame [18:15] or pc bios, if you will [18:15] you're right, man says d is cdrom [18:16] thanks for catching that [18:16] np [18:16] now I need to figure out my systemd unit and I'm all set [18:16] for systemd I'm not your guy :) [18:17] it's actually not too bad to run qemu on its only, only took me 5 days and harassing half of the ppl in here :P [18:17] its own* [18:17] but there's no provision to start instances at boot so need to write your own glue [18:17] this looks promising tho: https://kissmyarch.de/archives/2014/02/28/qemu_systemd_service/index.html [18:18] at least that's how that guy did it and it seem a reasonably clean soluition [18:18] stop/reset on socket works like a charm, already using that to interact with daemonized instance [18:18] have you tried libvirt and virt-manager? [18:18] you can use virt-manager to talk to remote qemu instances even, via ssh [18:18] I have and I'm against violence [18:18] :P [18:18] lol [18:19] I just remembered it because libvit will start vms on boot if you want it to, and it can also be used remotely in a headless scenario [18:20] yeah, I read about that, I gave it a try and decided not to go down that path, had too many problems right off the bat and since I'm mostly running lxd to make what looked like a big investment to figure it out didn't seem warranted [18:21] basically I like magic as long as I understand the spell [18:21] so I didn't trust myself to just point and click, not that it worked tho, couldn't get it to use my existing bridge for one [18:22] I'm not saying it's bad, don't get me wrong, I hear lots of ppl happily use it in prod [18:22] just didn't seem to be a good fit for me/what I'm doing even tho there was some upfront work involved this way too [18:23] sure [18:37] anothe trying I'm trying to figure out, so far I always used -drive file=.... as I was dealing with image files [18:38] but now I'd like to pass a partition of a disk to be used by KVM as a data mountpoint [18:38] file= doesn't seem to be the right thing, but I can't google out what enchantment exactly I should use [18:40] or maybe it is, just found something where they use lvm vols straight in file= syntax [18:47] \0/ worked [18:48] whups, cyclop