[00:57] <m4xx> evening all
[00:57] <m4xx> every time i reboot my 16.04 server my raid needs to resync
[00:58] <m4xx> i found on google people have had success adding "mdadm --wait-clean --scan" to their shutdown script but i don't know where to add it
[01:38] <cncr04s> how is the raid setup?
[01:39] <cncr04s> is gour grub bootfile up to date?
[01:42] <m4xx> to tell the truth i'm not sure. i set up the raid in the post conf setup before i installed the OS
[01:42] <m4xx> it asked me if i wanted to activate it during the install
[01:43] <m4xx> @cncr04s -^
[01:47] <cncr04s> so,
[01:47] <cncr04s> you did not manully setup software raid in ubuntu installation?
[01:47] <cncr04s> is it via a raid card or motherboard? is it intel?
[01:48] <m4xx> yes, via motherboard
[01:48] <m4xx> it is intel
[01:48] <cncr04s> then you are likley running into the intel fake raid issue
[01:48] <cncr04s> an issue i faced
[01:48] <cncr04s> i solved it by running software raid, setup via the installation
[01:48] <m4xx> i did run update-initramfs -k all -u
[01:49] <cncr04s> what particular motherboard is it
[01:49] <m4xx> not sure, i got the box from a friend. it's a zotac of some sort
[01:50] <cncr04s> 99% chance its fake raid if its a desktop pc motherboard
[01:50] <cncr04s> unless 16.04 included some changes to support it, it won't work
[01:50] <m4xx> would it be possible to re-create with out having to kill my data?
[01:51] <cncr04s> you can't recreate it without trashing what already exists on the disk
[01:54] <cncr04s> If you don't have a dedicated raid chip via a addon card or on the mobo, then its going to be os software managed any how. You might as well just go with the ubuntu softraid without dealing with the extra interaction with the mobo isw raid stuff.
[01:55] <cncr04s> that's my opinion, it worked for me. up to you though
[01:55] <m4xx> does the server install walk you through it? i'm kind of a noob
[01:56] <cncr04s> I don't know if a noob can do it. It has options for you to set it up in the installation. the part where it asks you to partition. there is an option to configure raid
[01:58] <cncr04s> searching gooogle/youtube for how to setup raid in ubuntu setup may yeild results, try that.
[05:49] <antony7777> i need help to clean up a server /boot ... using apt-get autoremove show this error msg: linux-image-extra-4.2.0-38-generic : Depends: linux-image-4.2.0-38-generic but it is not installed
[05:51] <antony7777> using: apt-get remove linux-image-extra-4.2.0-38-generic .. shows message: linux-image-extra-4.2.0-42-generic : Depends: linux-image-4.2.0-42-generic but it is not going to be installed  linux-image-generic : Depends: linux-image-4.2.0-42-generic but it is not going to be installed                        Recommends: thermald but it is not going to be installed
[05:51] <antony7777> the /boot is full, but I can't (don't know how) to free it
[05:52] <antony7777> my uname-r is: 4.2.0-16-generic
[05:53] <antony7777> so I'm guessing that I had a linux kernel update failed..?
[05:53] <antony7777> can I just go to /boot and delete the folders of newer kernels .. newer than the one I'm using, ie 4.2.0-16-generic .. will that be safe?
[05:55] <Ben64> don't delete newer ones
[06:00] <antony7777> the problem is I'm actually using older kernel.. 4.2.0-16 .. all other kernels in /boot are newer than 4.2.0-16, eg 4.2.0-27, 4.2.0-30 .. etc
[06:02] <Ben64> ew
[06:02] <Ben64> so do some "sudo dpkg --force-all -P <package>" and then run update and dist-upgrade and autoremove
[06:03] <antony7777> like: sudo dpkg --force-all -P linux-headers-4.2.0-42  ?
[06:11] <Ben64> don't do the newest one
[06:22] <ubuntu_> I am wondering where plymonth come into the linux boot loading ? I know bios runs code, which runs bootloader like grub ,lilo,..etc, which then load kernel , init.d... But whats the point of plymonth if its running after the kernel is loaded an operational?
[06:23] <ubuntu_> what i want to know is where plymonth falls in i know it has to be before the gdm login but after the bootloader code
[06:25] <ubuntu_> And with grub2 you can have the grub menu and graphics screens so its like an in between graphic to the grub and the gdb login page. But is there anything useful happening between that time or is it just for eye candy
[06:25] <Ben64> head -n5 /etc/init/plymouth.conf
[06:28] <ubuntu_> so is this before or after init starts?
[06:28] <Ben64> everything is after init
[06:29] <ubuntu_> Ok then whats happening between plymonth and  gdm login ?
[06:29] <Ben64> everything
[06:30] <ubuntu_> but if init starts all the programs then plymonth is just wasting time to go directly to gdm
[06:30] <ubuntu_> I guess i am wondering on what is being set up or running when plymonth is running
[06:30] <Ben64> what are you trying to get at
[06:31] <ubuntu_> Like if init already has all the programs mostly running when plymonth , usplash,..etc  is then run  one wouldn't see a lag if he went directly to gdm login
[06:33] <Ben64> no
[06:34] <ubuntu_> if init is just running plymonth as a process if plymonth is the last process before gdm login is kicked off then there really isn't to much point in having it . If its the starting process init kicks off i can see the nice screen while other process are getting started
[06:35] <Ben64> whats your point
[06:35] <ubuntu_> is it the former or the latter i suppose one can configure init to start the process in  different order to some extent but never looked into that much indepth yet
[06:37] <ubuntu_> My point is if its at the begining of init you looking at eye candy that actually is there because your still waiting for process to be started/configurations to be set. If it is at the end right at the point of gdm login then the eye candy is  just there wasting time
[06:37] <ubuntu_> to get to gdm
[06:37] <ubuntu_> login
[06:37] <Ben64> ok?
[06:38] <Ben64> i feel like you're leading to an actual question
[06:39] <ubuntu_> Well can you have a splash screen like plymonth at the grub level before one boots a kernel. Between the grub menu and the loading of the kernel/initramfs stuff
[06:39] <ubuntu_> ?
[06:44] <ubuntu_> I guess the grub menu is the splash screen because the rest of kernel loading and ramdisk would take  to get to init to plymonth neglatible time
[06:45] <ubuntu_> so i guess never mind kind of thought thru this one
[06:55] <ubuntu_> On an unrelated question is about installers like ubiquity , debian-installer , Anaconda  to name a few how are these set to only run  on an install cd and not on a live-cd by default is it part of init.d process folder the one adds the installer to boot up from
[06:56] <ubuntu_> because why i am at my live-cd or a computer on ubuntu does matter live or not ... i can  issue ubiquity at the command prompt and get the installer to start
[06:59] <ubuntu_> So i am just trying to figure out what the best way is to have it always start on boot another words make this a install cd or usb... i was thinking just to add it to the init.d process that init starts or one of its subprocess start but not sure on the best place to put it or what order it should start on ... probably towards the end of init spawning
[07:02] <ubuntu_> More i think about it .. it will need x11 so it has to be a last process kind of thing
[07:03] <ubuntu_> Note i could do a text based install or preseed but either way i know how this all works it just where to place these programs when booting
[07:10] <InnerCode> Hi, I've a strange problem on an Ubuntu 16.04.1 host with lxc/d 2. I've created a bridge on my eth0 connection and ran lxd init to select the bridge as default connection. After that I launched an Ubuntu 16.04 container and it gets his IP address from my DHCP server. I changed the /etc/network/interfaces file from the container to set it a fixed IP address. After stopping starting the container
[07:11] <InnerCode> it have 2 IP addresses. The DHCP one and the fixed one. I turned my DHCP server off on my network and after a reboot from the container the virtual interface won't go up. So, how do I set the IP address of a container correctly or how to remove the DHCP one?
[08:31] <negev> hi, i'm having issues with the dovecot apparmor profiles, specifically:    apparmor="ALLOWED" operation="sendmsg" info="Failed name lookup - disconnected path" error=-13 profile="/usr/lib/dovecot/log" name="run/systemd/journal/dev-log" pid=2603 comm="log" requested_mask="w" denied_mask="w" fsuid=0 ouid=0
[08:31] <negev> writes to run/systemd/journal/dev-log are allowed in the base abstraction which is included in the /usr/lib/dovecot/log profile
[08:31] <negev> but for some reason it doesn't work
[08:34] <InnerCode> I found it! I had to comment out the line 'source /etc/network/interfaces.d/*.cfg' in the interface file in the container. It had a cfg file that enables the DHCP for the eth0 interface. Mystery solved.
[10:50] <m4xx> so i've rebuilt my raid using this guide: http://www.howtogeek.com/51873/how-to-setup-software-raid-for-a-simple-file-server-on-ubuntu/
[10:51] <m4xx> but it wont come up after boot
[11:20] <cpaelzer> smb: puah - if it is a good one would you like to make some advertisement for the AC I sometimes see in your office so I can buy one ... :-) ?
[11:45] <smb> cpaelzer, its okayish. the main problem is to get the heat blown to the outside ... and having insect nets there makes that hard. plus its loud
[11:49] <cpaelzer> smb: hmm I have space incl windows in the basement below me for "loud" and "air out" - I need to properly investigate that
[11:50] <cpaelzer> smb: ignoring my favor of light I might also just work down there in the few hot days we get - if I would not need to move sceens and all that ...
[11:53] <smb> cpaelzer, yeah. I kind of get along by letting it run while I am not in the room and try to keep it dark the rest of the time
[13:16] <negev> hi, is there anyone around who knows apparmor really well?
[13:16] <patdk-wk> !ask
[13:17] <negev> appreciated, but i've done that several times over a few days and didn't get an answer so :P
[13:17] <patdk-wk> what makes you think anything will improve by asking like you just did?
[13:17] <patdk-wk> if they didn't know then, why would they know know and have enough time to stick around for you to respond that they did?
[13:18] <negev> 1) why does the base abstraction that grants write access to systemd dev-log not work?    2) is there a way to handle wrapper scripts that doesn't involve granting execute of /bin/bash to all child processes?
[13:18] <negev> patdk-wk: someone told me one of the apparmor devs hangs here sometimes
[13:18] <patdk-wk> so? I don't see what that has to do with any of the above
[13:19] <patdk-wk> you are wasting his and your time, if he responds, just by asking that question, without asking what you wanted to know
[13:19] <negev> see above, i asked two specific questions
[13:23] <Phrygian> I've just purchased an ubuntu server to run a server for a game, but I'm having some issues. I have to run one node js program followed by another, but the first program takes up the console so when i run a command it goes straight into the first program instead of the terminal. How do i avoid this?
[13:30] <andol> Phrygian: The quick-fix is probably to use something like screen or tmux, and give each program its own terminal. Also, by using screen/tmux it also becomes possible to detach from those terminals, and have the programs continue running in the background.
[13:31] <andol> Phrygian: Yet, the real solution is probably to setup init scripts (systemd units) and threat them as long running background services, assuming that that is what they are?
[13:38] <Phrygian> Yeah, I think i'm just going to combine the programs though, that seems like a more appropriate solution
[14:08] <negev> part
[15:41] <negev> hi, if i run the same application with different configs, is there a way to have a different apparmor profile for each config?
[15:45] <jaguardown> Hi all, do-release-upgrade on 14.04.2 says there's no new release found even after updating repos.
[15:45] <jaguardown> Did 16.04.1 not get released on the 21st?
[15:46] <negev> jaguardown: it did, but you have to update the metafile manually because it's not been updated yet
[15:46] <negev> i dunno why
[15:46] <jaguardown> oh okay, thanks
[16:09] <teward> jaguardown: they didn't open the release path yet
[16:09] <teward> negev: ^
[16:09] <teward> negev: bug squishing and other things need to happen first
[16:09] <teward> which is why it's not open yet
[16:09] <teward> at least, last i heard
[16:09] <negev> makes sense
[16:09] <negev> i've filed an apparmor bug already :P
[16:14] <negev> is there a way to customise apparmor profiles with environment variables when an application starts up?
[16:18] <negev> manpage for apparmor on ubuntu.com is a 404: http://manpages.ubuntu.com/manpages/xenia/en/man7/apparmor.7.html
[16:18] <negev> linked from: https://wiki.ubuntu.com/AppArmor
[16:20] <rbasak> negev: thanks, fixed.
[16:20] <rbasak> negev: also it's a wiki, so please do feel free to fix any other mistakes!
[16:21] <negev> maybe when i learn how to use apparmor properly, at the moment i'm having trouble working out if my issues are bad config or bugs
[16:23] <jaguardown> teward: Thank you, I was just about to ask how to manually update that metafile.
[16:23] <jaguardown> But I'll just wiat
[16:23] <jaguardown> wait*
[16:23] <negev> the upgrade went pretty much fine for me
[16:26] <teward> negev: YMMV
[16:26] <teward> so :P
[16:27] <teward> jaguardown: yeah I'd just wait until it's opened up
[16:27] <negev> if it's still buggy why was it released?
[17:18] <ubuntu_> installers like ubiquity , debian-installer , Anaconda  to name a few how are these set to only run  on an install cd and not on a live-cd by default is it part of init.d process folder the one adds the installer to boot up from ??? ..... and i am trying to figure out if you can have grub directly boot the x11 /graphics services , and the installer program without first booting into an OS or ramdisk. I know x11 is userspace
[17:18] <ubuntu_> program
[17:18] <ubuntu_>   so there would have to be an equivalent grub mod for that which supports what the installer program needs....
[17:38] <sarnold> ubuntu_: booting straight into graphics is slightly difficult; the plymouth program can give a better illusion that the system is running graphically earlier than X11 really starts.. (I think. at least that used to be the case.)
[17:40] <ubuntu_> plymonth still needs x11 /Xorg to run
[17:41] <ubuntu_> And not talking about a splash screen now i am talking about where to place the installer gui programs  when creating an install distro from a live one
[17:41] <JanC> Ubuntu disc images are both
[17:43] <ubuntu_> are you talking about the media the distro will be in like iso9660 cd/dvd as opposed to usb or floppy img  i don't get you
[17:43] <ubuntu_> why would that even matter
[17:43] <JanC> and it happens in init configuration/scripts and/or in xinit/xsession and/or in similar places; grub (or other bootloaders) can add options to the kernel commandline (which the script can read from /proc/cmdline )
[17:45] <ubuntu_> I can run  a splash screen like plymonth ,usplash or any other and i can run any gui or text based installer in userspace at a terminal what i want to know is where typically you place these programs to startup
[17:46] <ubuntu_> Order matters so should i uses the runlevel /etc folders
[17:46] <ubuntu_> Or is there a better way to most people do it
[17:46] <JanC> that's going to be different for every distro, right?  but for sure an init script/config (sysvinit/upstart/systemd/...) will be needed
[17:47] <JanC> best is to look at some distros & see how they all do it?
[17:49] <JanC> (nowadays one could also write an installer that runs under UEFI instead, I guess)
[17:50] <ubuntu_> Ok so its going to be dependent on what the system is using systemd or other things and looking up how to start process in a certain order ok. Is there any way if one is doing a text based install to just have grub , the installer programs it uses , and the installer on an iso image /usb img  with no OS or kernel. Because in theory if your doing text based you don't need user land only the mounting programs , making file sys
[17:50] <ubuntu_> tem programs  , and the grub drivers i would image
[17:53] <ubuntu_> And if you throw in a grub network PXE boot i would imagine you only need grub for your install usb or cd/dve nothing else on the media other then the boot/grub folder
[17:58] <JanC> ubuntu_: like I said, nowadays you don't even need GRUB, you could implement it as an EFI/UEFI application
[18:00] <ubuntu_> How does that work you still need a bootloader with PXE boot unless your talking about a built in bios PXE boot option that one can set where the tftp site is on the network
[18:00] <JanC> UEFI has (or can have) network drivers, graphics drivers, etc.
[18:00] <JanC> it's a basic OS on itself
[18:00] <ubuntu_> at the bios level
[18:01] <JanC> you can even run Python on UEFI
[18:01] <ubuntu_> if its at the bios level then the bios has become a mini os itself
[18:01] <ubuntu_> in the firmware
[18:02] <JanC> "BIOS" was the name of the old firmwares based on the original IBM PC, UEFI replaces it
[18:02] <JanC> but yes, it's sort of a mini OS
[18:03] <JanC> the UEFI firmware in my desktop has a GUI for configuration, can update itself over the network/internet, etc.
[18:04] <ubuntu_> got it so kind of just wondering if i didn't want to uses UEFI and wanted to uses some type of physical media to give a person like a cd/dvd or a usb  what would be the minimal stuff i need could i do it with just a bootloader accessing a network resource i.e just have /boot/grub
[18:04] <ubuntu_> I think that was possible but never tried it
[18:04] <trippeh_> on apple computers with UEFI you can even restore your computer from icloud from the UEFI/BIOS
[18:04] <JanC> UEFI can boot from CD?
[18:04] <ubuntu_> Can grub do PXE
[18:05] <ubuntu_> Its for an old computer that doesn't have UEFI or support it
[18:07] <ubuntu_> I could make the traditional install or live cd/dvd/usb but was curious on if i could have the iso , img on the network and just give him a cd/dvd/usb that has grub on it thats it that does the PXE boot to install the image from the network. Its to bad that UEFI is not supported
[18:09] <ubuntu_> I just wanted to save space and not but it on the cd/dvd/usb and the fact that he has 10 old computers that don't have UEFI this is just an example not the really thing just a made up example.
[18:15] <JanC> you can use PXE from UEFI/BIOS or from GRUB, I think
[18:15] <JanC> (not sure if GRUB2 has PXE support, but GRUB certainly did)
[18:17] <JanC> http://wiki.osdev.org/Diskless_Booting#The_gPXE_.2B_GRUB_Way
[18:17] <ubuntu_> Ok cool so then in that case all one would need on his media is a /boot/grub folder or whatever folder he wants to call it for grub but essentially just grub access a network image
[18:18] <JanC> that osdev site might have more useful info for whatever you are trying to do  :)
[18:18] <ubuntu_> about the network image for old computers i would imagine it would have to be in a folder form not a cd/dvd iso9660 600MB to 4GB because one would need the ramdisk space to store it in?
[18:20] <ubuntu_> How typically is PXE booting done when one has limited memory does it like temporarilly copy to an external hdd by default or something?
[18:22] <ubuntu_> I mean i get if the iso or img file can fit completely in memory then one can uses PXE from grub to install retrieve it into memory and install it from there but if one doesn't have the ramdisk space i would imagine you cann't do the PXE boot unless it can be specified to uses a swap partition on some secondary harddrive for temporary storage when installing
[18:23] <ubuntu_> Typically now a days everybody has enough memory not to have to worry about not fitting in memory the image but back in the day when one only had say 128 or 256 MB how could the image fit.
[18:25] <ubuntu_> I would imagine since your installing the os one would always have available a secondary storage that did have enough room but then PXE grub program would have to uses some tmp space for it
[18:25] <ubuntu_> to store the iso image and load only the subset of files
[18:26] <ubuntu_> Or maybe the tftp site only has the essential files of the iso placed on the network for older computers... and for newer ones they can just uses full iso or img images from the network... curious how this uses to work from how it works today
[18:32] <patdk-wk> heh
[18:32] <patdk-wk> I just iscsi boot
[18:48] <ubuntu_> Ok cool iscsi for the initator connection to the target by the IQN number . I get it but what devices can the target uses for the initator that show up in device is it just a virtual disk on the target  or can it be the actual part or whole physical HDD connected to the target like a usb drive or a slave HDD
[18:48] <ubuntu_> Because i just say stuff on the target using a virtual disk to provide the initator
[18:49] <ubuntu_> so can it also uses non virtual HDD , cd rw , usb ..etc device or do most people uses a virtual HDD on the target for the initator iscsi machine?
[18:53] <ubuntu_> I get the difference between using iscsi SAN technology as opposed to using NAS storage technology. But typically what does the iscsi media uses for providing storage for the initator?
[18:56] <sarnold> iscsi initiator may be booted via pxe or a sata DOM mini-ssd or USB mass storage device
[18:58] <ubuntu_> Ok at that point its just like having the network harddrive or cluster locally to you when booting ( kind of affect)
[18:59] <ubuntu_> Is that currently the best way for providing network booting/installs for large companies?
[19:03] <patdk-wk> it depends on the needs
[19:03] <patdk-wk> after all that, still no idea what your needs are
[19:03] <patdk-wk> my needs where simple
[19:03] <patdk-wk> one iscsi volume
[19:03] <patdk-wk> the iscsi volume is cloned for each workstation at boot
[19:03] <patdk-wk> and destroyed on shutdown
[19:05] <ubuntu_> iscsi can uses any type of physical storage device right like cd/dvd/usb/floppy/nas/....etc device or is it just fix to hardware that uses the scsi commands so maybe not ATAPI cd/dvd not sure
[19:06] <ubuntu_> ?
[19:06] <patdk-wk> heh?
[19:06] <patdk-wk> iscsi is limited to scsi
[19:06] <patdk-wk> not sure why that wouldn't include cd/dvd/usb/floppy
[19:06] <ubuntu_> so what device now a days are not scsi based
[19:06] <patdk-wk> yes, you cannot run nas over iscsi
[19:07] <patdk-wk> well, usb isn't scsi normally
[19:07] <ubuntu_> just trying to figure out what hardware it won't work on
[19:07] <patdk-wk> my first cd was scsi
[19:07] <patdk-wk> but not sure what any of this has to do with anything
[19:07] <ubuntu_> O so no usb device darn  you have to uses NAS in that case for the usb devices
[19:07] <patdk-wk> why?
[19:08] <patdk-wk> but then, if your using usb, you did something seriously wrong
[19:08] <ubuntu_> because how else would you have a computer access a usb HDD on another computer  thru then network
[19:08] <patdk-wk> you don't
[19:08] <patdk-wk> doing so is foolish and unreiable
[19:08] <patdk-wk> usb it not known for it's stability
[19:09] <patdk-wk> but who said you cannot use usb disks over iscsi?
[19:10] <ubuntu_> Also can iscsi uses virtual HDD or virtual cd/dvd drives instead like if one  has vmware  HDD files on his iscsi target device instead of physical devices
[19:11] <ubuntu_> i.e virtual scsi device instead of physical scsi devices
[19:13] <ubuntu_> Because if it can that would be a really good way for business to have many peoples computers just have a different virtual scsi HDD on the network i.e share space on the target scsi  as virtual files that are shown to the initator as really physical devices
[19:14] <JanC> that would depend on the iSCSI target (NAS/SAN/server) you use, I guess
[19:14] <JanC> iSCSI itself is just a protocol
[19:16] <ubuntu_> Humm curious partitioning the physical device like one partition for each person with 1000 people then setting quotas seems impratical but using 1000 virtual HDD files  that each of the client initator machines uses seems more practical
[19:18] <ubuntu_> I guess one could do the traditional way and have there local laptop or desktop HDD just for there files and the NAS or SAN as having there os partition but your still going to need partitions for each of the 1000 if your not using virtual HDD
[19:19] <ubuntu_> Or one could uses Active directories instead
[19:32] <ubuntu_> Just thinking out loud since if one had one iscsi with windows 10 os on it and all the clients just PXE boot to the iscsi with there whole local hard drive as  just file and there program storage then would the performance halt for 10000 people using this same iscsi target to log in
[19:34] <ubuntu_> seems it would so kind of wondering the best way to set up pxe / UEFI booting   to iscsi target when it comes to 1000 s of computer users... i would think you would have to increase the different iscsi targets which would kind of increase the number of computers or at least scsi devices
[19:41] <ubuntu_> Also sinces iscsi is a SAN i would imagine you could set this up with a cluster of scsi devices manybe thats how they do it not sure
[19:42] <ubuntu_> This wouldn't seem like it would help in different os's images only in performance , load balanceing , and failover raid backup,..etc stuff so kind of still confused