[03:27] <masuberu> hi
[03:27] <masuberu> has Ubuntu a tool or a way to restore the / file permission?
[03:40] <sarnold> masuberu: if you've changed / to something that then forbids you from running commands, you may not have many options except rebooting into a rescue system
[03:41] <masuberu> ok
[03:41] <sarnold> masuberu: maybe you can run programs via ../../bin/chmod
[03:41] <masuberu> you mean boot from original cd and restore it from there?
[03:41] <sarnold> yeah that should do it
[03:41] <sarnold> what di dyou change it to? :) and how? :)
[03:42] <masuberu> is there an option to fix it on the installation menu on the boot cd?
[03:42] <masuberu> mate, I did the most stupid thing on earth
[03:42] <masuberu> I run sudo chmod -R 777 / nvme
[03:42] <sarnold> ahhhhhhh
[03:42] <masuberu> instead of sudo run chmod -R /nvme
[03:42] <sarnold> so you've got a lot more than just the permissions on / to fix
[03:42] <masuberu> yes
[03:42] <masuberu> basically everything
[03:43] <masuberu> what would you recommend?
[03:43] <sarnold> after a good long cry?
[03:43] <masuberu> is reinstall my only solution?
[03:44] <sarnold> that's the best solution, but you might be able to get away with comparing against another system
[03:44] <sarnold> .. and then manually fixing everything. that doesn't sound like fun though.
[03:45] <masuberu> I thought ubuntu installation cd could fix it
[03:45] <masuberu> choosing an option from the installation menu?
[03:45] <sarnold> afaik debian/ubuntu doesn't have any equivalent of other systems's "expected permissions" tools, and I don't think dpkg has any way to report which permissions are wrong and what they -ought- to be :(
[03:45] <masuberu> hum ok
[03:46] <masuberu> now my problem is that I can't ssh into it ...
[03:46] <sarnold> .. because the permissions are too permissive. oooof.
[03:46] <masuberu> I go to terminal and I can ping outside
[03:46] <masuberu> the ip of the machine is same as before so no network issues
[03:47] <masuberu> does it ring any bell?
[03:47] <sarnold> yeah, ssh is really picky about allowed permissions on its files
[03:48] <sarnold> try first chmod 755 / /etc /etc/ssh /home ~ ; chmod 700 ~/.ssh
[03:50] <masuberu> ok...
[03:50] <masuberu> hasn't complain
[03:50] <masuberu> shall I restart the machine?
[03:55] <sarnold> depends; if you want to just re-install, that's not a bad option. if you want to try to fix it, just keep changing permissions on it until it's fixed.. _then_ reboot
[06:01] <masuberu> i ended up reinstalling
[06:01] <masuberu> :___(
[09:18] <phoenix_> hello
[09:19] <phoenix_> i have configured postfix
[09:19] <phoenix_> when i try echo "Test mail from postfix" | mail -s "Test Postfix" you@example.com
[09:19] <phoenix_> echo "Test mail from postfix" | mail -s "Test Postfix" dahlia2.co2016@gmail.com
[09:19] <phoenix_> i get this error
[09:19] <phoenix_> Can't canonicalize "/home/dahlia/Maildir"
[09:20] <phoenix_> Failed to save message in "/home/dahlia/Maildir/sent" - message not sent
[09:20] <phoenix_> "/home/dahlia/dead.letter" 6/140
[09:20] <phoenix_> what should i do ?
[09:32] <phoenix_> is there someone?
[09:38] <hateball> !patience | phoenix_
[09:38] <hateball> (I don't know)
[09:39] <hateball> phoenix_: Permissions, perhaps? Also there is #postfix
[09:40] <phoenix_> #postfix :Cannot send to channel
[09:45] <hateball> !register | phoenix_
[09:45] <phoenix_> hateball, i have a domain name dahliaco.com and it is configured on a server i need to my customers recieve emails from info@dahliaco.com
[09:46] <hateball> phoenix_: I don't know anything about postfix so there's no need to direct towards me :)
[09:46] <phoenix_> i dont know how can i do that
[10:35] <phoenix_> omg
[11:39] <pirx> when i install 14.04 i am asked about the Location (of the server i presume). anyone have an idea of where this is saved? or is it seen in e.g. the timezone only?
[12:14] <Xin> hey all whats an easy to install web proxy?
[12:15] <hateball> Xin: nginx
[12:15] <hateball> Xin: or what do you mean by a proxy? does it need to cache things and so on?
[12:15] <hateball> !info squid
[12:16] <Xin> I dont really care about caching
[12:16] <Xin> ill give squid a crack though, thats what ive always heard about
[12:16] <Xin> thanks
[12:25] <Xin> lordy
[12:25] <Xin> this config is complex lol
[12:49] <lamont> pirx: /etc/timezone
[12:50] <foormea> hi. i'm working on a headless server and i'm about to change its mobo. do i have a way to predict the network interface name with the new mobo? with ethX renamed in enXXXX i'm a bit wary of screwing up my config (and i have no access to console or serial on that headless server)
[12:50] <patdk-wk> yes
[12:50] <patdk-wk> or add a udev rule
[12:52] <foormea> mh, could you explain further how to predict the name, or how to add the udev rule, or point me to some doc? :) sorry i'm a bit clueless right now :/
[12:54] <ogra_> the location also makes apt pick the closest server for your sources.list
[13:00] <foormea> patdk-wk, got it: https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/    thanks for the hint!
[13:02] <pirx> lamont: thanks!
[13:33] <foormea> back to my question about predictable network interface naming: can i somehow have both the predictable (enpXsY...) AND the former standard (ethX) pointing to the same device?
[13:34] <foormea> this way i'd be able to configure my eth0 during the motherboard change
[13:34] <foormea> as it's gonna be tricky to know the "predictable" name for that specific slot on that new mobo :/
[13:34] <ogra_> you mean the unpredictable names taht everyone knew from the top of his head ? :)
[13:35] <ogra_> you can drop "predictable names" completely by setting net.ifname=0 on your kernel commandline
[13:35] <ogra_> that will switch bac to the old behaviour
[13:35] <Odd_Bloke> *net.ifnames=0
[13:36] <foormea> well i don't really wanna do that: i don't wanna use an older standard. i'd like to have both ethX and enpXsY for a bit, so that i'll know after mobo change the new iface name
[13:36] <foormea> if i disable it altogether, i won't know the new name under predictable scheme
[13:36] <ogra_> ah, thanks for the correction (and sorry)
[13:37] <foormea> i'm thinking of having a script run at bootup that dumps "ip a" into a file, boot with new mobo
[13:37] <foormea> reboot with old mobo and see the file
[13:37] <foormea> modify config accordingly...
[13:37] <foormea> that's a bit overkill
[13:38] <foormea> but i have no serial/screen/keyboard AND i have no usb HDD reader
[13:38] <Odd_Bloke> foormea: No DHCP server?
[13:38] <foormea> i have a dhcp running on the network yeah
[13:38] <foormea> bt
[13:38] <foormea> but
[13:38] <foormea> if that iface is not told in /etc/network/interfaces to come up... it won't, will it?
[13:39] <foormea> and the thing is, i need to be EXACT in what i'm doing, or i lose my server until i get a screen/keyboard/etc for it :/
[13:42] <Odd_Bloke> foormea: Well, you can specify configuration for both eth0 and $new_name and still have the option to switch back to the old mobo, right?
[13:42] <foormea> any thoughts, anything obvious i haven't thought of (apart from reverting to older ethX naming scheme)?
[13:42] <Odd_Bloke> https://forums.opensuse.org/showthread.php/493546-How-to-predict-predictable-network-interface-names <-- there's a script there which purports to tell you the predictable name
[13:42] <foormea> yeeeeeeeeeh
[13:42] <Odd_Bloke> (Note that I had to change the path of lspci, but it worked for me)
[13:42] <foormea> ohhhh let me see
[13:43] <foormea> in that case i'd revert to ethX scheme, change mobo, check the "predictable" name with that script, revert config to predictable
[13:43] <Odd_Bloke> foormea: Right, because you don't actually have the system booted with the new hardware.
[13:43] <foormea> yep
[13:44] <foormea> aright very good, very helpful that link you just gave me
[13:44] <foormea> anyway the background story is: i'm sort of scavenging hardware
[13:44] <foormea> and just got a hold of a newer mobo/cpu than what i had
[13:45] <foormea> argh grrr i hope virtualisation is enabled in that mobo's bios...
[13:46] <Odd_Bloke> foormea: TBH, you might be best off scavenging a keyboard/monitor and doing this another day. ;)
[13:47] <foormea> yeah perhaps you're right, but i like a challenged
[13:47] <foormea> challenge
[13:48] <foormea> well thanks for the tips, i'll give it a thought and let you know how it went if i end up doing it
[13:50] <Tegu> I had to move a 16.04 usb installation to another machine and had the same naming issue. fortunately, I could connect the other machine to a monitor and check the name there. but I kinda wished it could be automagic :)
[13:56] <foormea> oh well i'll go ahead for that solution: revert to older naming scheme, get the new name, revert to newer naming scheme
[13:57] <foormea> worst case it screw it up and won't be able to use it for a few days -- found a keyboard, but i'll need to find a vga cable which i can find in 3 days
[14:00] <foormea> so i add net.ifnames=0 to end of line "linux   /boot/vmlinuz-4.4.0-28-generic root=UUID=8c3f5192-2348-45f9-ad7f-83360f82589b ro" in my /boot/grub/grub.cfg ?
[14:01] <foormea> or... http://askubuntu.com/questions/19486/how-do-i-add-a-kernel-boot-parameter might be better
[14:06] <Dulcin> Hi does ubuntu have a default mail server enabled? I just uninstalled postfix, i dont want my testserver to send out mails
[14:06] <Dulcin> I probably installed postfix myself I dont remember, but is there anything else I should worry about?
[14:06] <Dulcin> I'm on Ubuntu 14.04 by the way
[14:18] <Xin> I need a romanian or estonian or somesuchhh
[14:19] <phoenix_> hello
[14:20] <phoenix_> i am configuring postfix
[14:20] <phoenix_> when i use  echo "This is the body of the email" | mail -s "This is the subject line" dahlia2.co2016@gmail.com
[14:20] <phoenix_> i get this error
[14:21] <phoenix_> Failed to save message in "/home/dahlia/Maildir/sent" - message not sent
[14:21] <phoenix_> "/home/dahlia/dead.letter" 6/159
[14:25] <Xin> anyone know a CHEAP vps provider in like, estonia/romania etc regions?
[14:28] <patdk-wk> lowendbox.com
[14:33] <yoink> KVM issues with 16.04 - basically I run a bunch of 14.04 virual machines (on a 16.04 host) without issue. However I can't seem to connect/setup 16.04 in KVM. Whether I update a working 14.04 VM to 16.04 or start with 16.04 from scratch, it won't get to the console to login.
[14:34] <yoink> I'm using libguestfs mostly - so virt-builder and virt-install and managing the vms through Virsh.
[14:35] <rbasak> yoink: uvtool works on 16.04 to start a 16.04 instance. AFAIK cloud images as published work with 16.04's KVM. So I wonder what virt-builder/virt-install are doing differently?
[14:36] <yoink> rbasak: I'll check with the #libguestfs folks. Other images work fine... maybe it's time to learn a new toolset! :)
[14:36] <yoink> Reading up on uvtool now. :)
[14:37] <rbasak> yoink: well the paths are certainly going to be very different. If there is a bug somewhere, it could be in libvirt for example, or just a change in behaviour that causes something unintended. My question is really more about identifying what that thing is to determine where the bug might be.
[14:37] <rbasak> I've just verified that uvtool successfully started:
[14:37] <rbasak> release=xenial arch=amd64 label=daily (20160707)
[14:38] <yoink> I hear that - I'm just running out of playtime. :)
[14:38] <rbasak> On a Xenial host (that isn't fully up to date probably, but was installed with Xenial freshly after release)
[14:39] <caribou> rharper: are you still planning on merging multipath-tools ?
[14:39] <yoink> rbasak: I have an idea... brb.
[14:45] <caribou> rbasak: rharper: we've just got a multipath-tools patch accepted by multipath-tools's upstream
[14:46] <caribou> rbasak: rharper: we'll do the submittodebian legwork but I thought that you might want to keep an eye on this one
[14:50] <caribou> rbasak: rharper: FYI, it's LP: #1570093
[15:00] <LaserAllan_> hey there, ive got a question, I have a couple of nfs folders that I want to be mounted at startup, should those commands be written with sudo in the rc.local script?
[15:04] <mdeslaur> LaserAllan_: no, add the mounts you want to /etc/fstab
[15:08] <LaserAllan_> mdeslaur: Oh, ok, what is rc.local then for launching applications and so on?, ive probably just missunderstood it all
[15:08] <LaserAllan_> but thanks
[15:08] <LaserAllan_> but should the nfs mount commands be added with sudo or is it automatic sudo on those?
[15:09] <mdeslaur> you don't directly add the commands, look at the file, you just need to add an entry to it
[15:09] <riz0n> Hello guys, I have an HP DL380 G3 server that I have installed Ubuntu LTS on. No GUI. Is there a way to change the system from loading the ATI Rage drivers to just loading Generic VGA drivers? The ATI drivers are causing my system to crash after so long of running.
[15:09] <mdeslaur> LaserAllan_: rc.local is run with root privileges, so anything you put in there doesn't need sudo
[15:17] <riz0n> Righ now, as it stands, when the system boots, eventually the system changes and the size of the letters on the screen get very, very small. I want standard VGA with the big letters. Any help would be appreciated!
[15:18] <LaserAllan_> mdeslaur: Is it the same with fstab?
[15:18] <LaserAllan_> Since i need to put the nfs mount commands in there
[15:18] <mdeslaur> LaserAllan_: the mounts in fstab are mounted as root, yes
[15:18] <mdeslaur> LaserAllan_: but you don't put commands in that file
[15:23] <LaserAllan_> mdeslaur: So where do I put them then?
[15:24] <mdeslaur> LaserAllan_: you put them in /etc/fstab, but not the commands. that file is a list, you just add what you want to the list.
[15:24] <mdeslaur> see here: http://manpages.ubuntu.com/manpages/xenial/en/man5/fstab.5.html
[15:26] <LaserAllan_> mdeslaur: Hmm, i have never really done this before so what do I put ther ethen to amek nfs folders mount on startup on my other host?
[15:27] <mdeslaur> something like "servername:dir /mntpoint nfs rw,hard,intr 0 0"
[15:27] <mdeslaur> see here: https://help.ubuntu.com/community/SettingUpNFSHowTo
[15:28] <LaserAllan_> mdeslaur: Thanks
[15:48] <riz0n> OK guys, I uncommented a line in the Grub config that put it in 640x480 then "update-grub" and it booted into 640x480. Any way to see what video drivers are loading and change them to just standard VGA? Or should it be standard VGA since there is no GUI?
[15:50] <nacc> !crosspost | riz0n
[15:51] <nacc> riz0n: what version of ubuntu?
[15:52] <riz0n> nacc, 14.04.21
[15:52] <riz0n> Oops 14.04
[15:52] <riz0n> whatever the latest rel is
[15:53] <nacc> riz0n: the latest release is 16.04 :)
[15:53] <riz0n> Well, the latest 14.04 rel ;)
[15:53] <riz0n> I want to say it's 14.04.3
[15:53] <riz0n> I'mts
[15:54] <teward> 14.04.4 is the latest point release :)
[15:54] <nacc> riz0n: ok, i think you should check what lspci says for your graphics device as to the driver in-use (and/or lshw)
[15:54] <riz0n> e that doesn't really matter, I'm on 14.04 LTS
[15:54]  * teward goes back to lurking
[15:54] <nacc> riz0n: it does matter, actually (given that you have different kernels, and stacks with each hwe stack, potentially)
[15:54] <riz0n> and my remote desktop is acting goofy
[15:54] <riz0n> This system is using the 4.2.0-41 kernel
[15:55] <riz0n> VGA compatible controller: Advanced Micro Devices, Inc. AMD/ATI Rage XL PCI rev 27
[15:57] <nacc> lscpi -vvv -s <pci address> of that device
[15:57] <nacc> riz0n: will tell you the kernel module in use
[16:00] <riz0n> Thanks so much guys. I know this DL380-G3 server is quite ancient, but it
[16:01] <riz0n> it's rock solid till the NMI PCI error happens (and I've traced it to the video). I hate to fill it full of buckshot over something simple I can easily fix. Plus I don't have the money to buy a new server yet.
[16:02] <riz0n> The error is "NMI: PCI system error (SERR) for reason b1 on CPU 0" Dazed and confused, but trying to continue
[16:05] <patdk-wk> hmm, I dunno I would bother messing with it much
[16:05] <patdk-wk> serr says the video card is bad
[16:05] <patdk-wk> throw in a pcie video card to replace it with
[16:05] <patdk-wk> and move on
[16:12] <rbasak> nacc: I'm a little confused as to why php7.0-mysql's {mysqlnd,mysqli,pdo_mysql}.so don't link to libmysqlclient at all. Could this be a bug? They did in PHP 5 in Trusty.
[16:12] <rbasak> Or are they statically linking by accident or something?
[16:19] <nacc> rbasak: looking
[16:20] <rbasak> nacc: no urgency. I'll ignore it for now.
[16:22] <rattking> Hello all, I would like to spawn a custom shell on one tty.. I am not sure how to do this since the switch to systemd. does anyone know where that is configured now? It use to be configured in /etc/init/tty?.conf
[16:23] <nacc> rattking: afaict, /etc/init/tty?.conf still exist with systemd
[16:23] <rattking> oh perhaps there is another reason its not working then.. thanks
[16:23] <rbasak> I don't think they're used though, are they?
[16:24] <rbasak> AIUI, /lib/systemd/system/getty* are what are used now.
[16:24] <nacc> unclear :/
[16:24] <rbasak> And you can override with /etc/systemd/system/
[16:24] <nacc> ah it does seem to be there now
[16:24] <nacc> rbasak: another case like our init.d discussion of confusing end-results :)
[16:25] <rbasak> Indeed.
[16:25] <rattking> heh
[16:25] <rbasak> It's a consequence of Debian supporting multiple init systems.
[16:26] <rattking> I will take a look around in /lib/systemd/system/getty* thanks for the tip
[16:31] <rattking> got it!
[16:34] <nacc> rattking: nice, congrats! :)
[16:46] <jrwren> anyone ever see this kind of systemd message? http://pastebin.ubuntu.com/18798467/  "Failed to foward REleased message"  it just spews to syslog every few seconds.
[16:58] <Toraxmalu> hi
[16:58] <Toraxmalu> hase someone experience with xen?
[16:58] <Toraxmalu> after installing xen-hypervisor-amd64 i was confronted with a boot-loop
[17:01] <PryMar56> Toraxmalu, is your system EFI?
[17:01] <Toraxmalu> yepp
[17:01] <Toraxmalu> it is
[17:04] <PryMar56> Toraxmalu,  dpkg -l xen-hyper*amd64 | grep -i efi
[17:04] <Toraxmalu> is there a known issue? I've read, that there are is the advise to switch back to BIOS, but that is simply not possible for my case
[17:04] <Toraxmalu> -are
[17:04] <PryMar56> maybe there is an efi boot blob in the *.deb
[17:04] <PryMar56> xen.efi
[17:05] <PryMar56> maybe not
[17:05] <Toraxmalu> err...
[17:05] <Toraxmalu> and that means?
[17:05] <PryMar56> Toraxmalu, upstream xen includes source for the xen.efi. Not sure is Xenial packages/builds it
[17:06] <Toraxmalu> okay - you're talking to a windows-power-user... and i understand, that xenial possible didn't compile that important stuff - do you have a link or something, so i can build that thing by myself?
[17:07] <Toraxmalu> (howto would be nive - sorry)
[17:20] <Toraxmalu> PryMar56 - I've found the Issue-List
[17:20] <Toraxmalu> it is simply not possible with UEFI with ubuntu 16.04
[17:21] <Toraxmalu> well - thank you for your efforts
[17:21] <Toraxmalu> and sorry for the distrubence