[01:00] <masuberu> I am deploying a software that runs on ubuntu 12.04 and I need to install git
[01:00] <masuberu> however apt-get can't find neither git nor git-core
[01:00] <masuberu> what can I do?
[01:03] <sarnold> masuberu: run "apt-cache policy git" .. it should show which repository it would install which version of the git package from
[01:04] <masuberu> sarnold: N: Unable to locate package git
[01:04] <sarnold> masuberu: strange. how about "apt-cache policy bash"? you've probably got that installed anyway, i wonder how it got there.. :)
[01:05] <masuberu> http://pastebin.com/raw/WwgZe90A
[01:07] <sarnold> masuberu: check your /etc/apt/sources.list file -- is the mirror there still valid?
[01:07] <masuberu> sarnold: I doubt, it doesn't even do apt-get update
[01:09] <ubuntu_> curious in the newer versions of ubuntu server i have come across when installing  MAAS install region or rack . I read about the hardware requirements i  know i don't have enough for those or openstack  more for large data centers then for small home networks. But curious what is the difference between region or rack installs?
[01:09] <ubuntu_> When would one do a region over a rack visa-versa
[01:10] <ubuntu_> As i understand it these are all cloud based computing technologies
[01:11] <masuberu> http://pastebin.com/raw/FQrFCnev
[01:12] <sarnold> masuberu: owwwww. I have to admit I wasn't expecting this one ;)
[01:12] <ubuntu_> try using a different repo location
[01:12] <sarnold> masuberu: are there nameservers in /etc/resolv.conf ? do they work?
[01:12] <tarpman> is the VM in question even connected to a network...? :)
[01:13] <ubuntu_> /etc/apt/source.list or uses the system control center or aptitude,...etc
[01:13] <masuberu> ups
[01:13] <sarnold> ubuntu_: there's some description of the rack controller vs region controller here https://maas.ubuntu.com/docs/install.html
[01:13] <masuberu> it is not resolving names ...
[01:13] <sarnold> tarpman: ha :) good question
[01:14] <masuberu> damn
[01:14] <masuberu> sorry my bad
[01:15] <ubuntu_> curious i know this is stupid but cloud computing seems to me like a virtual remote desktop  session to many different servers. Like a hyper-v on many different servers
[01:15] <sarnold> ubuntu_: most 'cloud' computers never have a desktop of any sort
[01:16] <ubuntu_> I know there is like paas, saas ,...etc distinctions of different clould services at different layers but seems as its all like a remote app , or rdp virtual session to different servers
[01:17] <ubuntu_> Right the cloud computer doesn't need a desktop just some hyper-v service base thing that clients with desktops connect to
[01:18] <sarnold> most of the time they don't have clients with desktops connecting to them either :)
[01:18] <ubuntu_> How is virtual remote desktop , remote apps,..etc anything different then what cloud is ... just cloud is using more computers but same prinicpals
[01:18] <sarnold> they run databases or webservers or fileservers or firewalls or streaming radio stations or antivirus scanners or irc bouncers or mathematica or openmpi or ... :)
[01:22] <ubuntu_> ya but at this level cloud is just like it was back in mainframe times just services on a remote computer that people could remote into uses... I get cloud makes it look pretty i guess thru web interfaces other stuff but non the less its not anything really new
[01:23] <sarnold> the basic "gist" of "cloud" is that there's an API that you can use to get new virtual machines or new storage devices or new IP addresses and attach them around
[01:23] <sarnold> ubuntu_: yes :)
[01:23] <sarnold> we've returned to the days of the "computing center", hehe
[01:24] <sarnold> except instead of a shell account on the university mainframe, it's an API endpoint that can spawn machines as users used to spawn processes...
[01:25] <ubuntu_> that would be like remote desktop into a virtual machine
[01:28] <sarnold> if that's a useful analogy for you, that's fine, but just be sure you know that most machines never run graphical programs and most never have any user interaction at all -- programs are installed and configured using tools like juju or chef or ansible, they run servers, and almost nothing ever actually interacts with humans...
[01:29] <ubuntu_> I get that
[01:29] <ubuntu_> What would you say grid computing and cloud computing differ... because i always confuse myself with the distinction of computing in the cloud as opposed to grid
[01:30] <ubuntu_> I would say they could be pretty much the same or over lap a lot
[01:30] <sarnold> 'grid' always feels like it runs a single program that handles jobs; jobs are distributed, run, and then collected; cloud usually installs services on virtual machines
[03:26] <Shambles> teward, sarnold -b worked with setfacl.  Everything looks clean again
[03:27] <Shambles> Really it wasn't super important but I didn't want some random ACL to remain on the root of my wifes share
[03:30] <ubuntu_> ya but what is cluster computing isn
[03:30] <ubuntu_> t
[03:31] <Shambles> I need to train myself to use the man command instead of --help.  Thanks for the assistance
[03:31] <ubuntu_> cluster nodes just mirror images of each other for load balancing and fail over
[03:31] <ubuntu_> because if thats the case cluster computing is just  load balancing computing not  computing different services on different nodes
[03:36] <ubuntu_> what i mean is to do different computing on different nodes one would have to uses more then one cluster
[03:37] <ubuntu_> if i am understanding cluster computing correct... i have never need clusters all that much and the computing part never understood the term... i do understand the  point to clustering for redundancy , and load balancing but for COMPUTING?
[04:07] <sarnold> Shambles: great :)
[04:07] <sarnold> ubuntu_: clustering is very rarely used for mirroring entire machines
[04:08] <ubuntu_> wait so a node is only  take pieces of software and mirroring them and the other part of the nodes can be very different
[04:09] <ubuntu_> I am looking at clustering as some kind of network raid between servers called nodes
[04:10] <ubuntu_> But if it can be pieces of servers that are mirror images with the rest distinct sections of server
[04:14] <ubuntu_> I would think they have to be mirror nodes since how else would fail over work... if a system goes down you need a mirror node to take over?
[04:14] <ubuntu_> For NLB clusters  i could see  nodes being not mirrors just the application being mirrored that are part of the NLB cluster ... just not sure
[04:26] <ubuntu_> Unless fail over clusters can be just setup as well to just mirror sections of the whole server like a database, web server,applicaton server,..etc. But to me when i think of a cluster it is  network raid for redundancy /backup and load balancing performance
[04:28] <sarnold> you can do both; database sharding is quite common, failover is less common
[04:37] <ubuntu_> fail over is like when they do network raid  / mirroring the server through out the different nodes in the cluster for  redundancy and uptime protection right?
[04:38] <ubuntu_> And NLB is more for just mirroring sections of servers like , application ,database,..etc
[04:38] <ubuntu_> if i understand you correct
[04:38] <ubuntu_> and thats the most common way's data centers uses them
[04:38] <temmi_hoo> clustering to solve computing problems is often done with splitting the problem to blocks handed out to large numbers of nodes
[04:39] <temmi_hoo> seti@home is a very large geographically distributed approach in clustering
[04:39] <ubuntu_> ok but then cluster computing is just say i was accessing data from a data base it may uses  node1 and node2 so its just computing that splits up the same task
[04:39] <sarnold> NLB?
[04:40] <ubuntu_> network load balancing
[04:41] <temmi_hoo> a database cluster is usually not a computational cluster
[04:42] <temmi_hoo> the database cluster might have a frontend machine or any node might act as a frontend, then the data is stored in a distributed fashion but the db user doesn't need to know where, upon making a db query the db cluster finds the data and hands it to the user
[04:42] <ubuntu_> what are these computational clusters your talking about .... because i can only see this type of computing between  different clusters not really different computations between nodes in a particular cluster
[04:43] <temmi_hoo> computational clusters such as any modern supercomputer in the last oh forty or so years
[04:43] <sarnold> check out this computational cluster :)  http://www.netlib.org/utk/people/JackDongarra/PAPERS/sunway-report-2016.pdf
[04:43] <temmi_hoo> also seti@home and the like
[04:43] <temmi_hoo> anytime you see weather report the forecast is computed on a clustered machine
[04:43] <ubuntu_> And curious is it possible to do iscsi  clusters like SAN clustering
[04:43] <temmi_hoo> it is
[04:44] <temmi_hoo> storage clustering can use iscsi or even nfs over ip as its communication media but in real performance oriented datacenters sas is used with specialized sas switching fabrics
[04:45] <ubuntu_> ok
[04:45] <ubuntu_> so cluster computing is just a form of distributive computing or splitting the database files pieces on seperate servers if its NLB based cluster... but if its fail over cluster you have to kind of have mirror images on nodes
[04:45] <temmi_hoo> not just database
[04:45] <ubuntu_> for your example
[04:46] <ubuntu_> Similar for other applications
[04:46] <temmi_hoo> it can be block storage or filesystem level storage or .... or raw number crunching
[04:46] <ubuntu_> So then how is this different then grid computing
[04:46] <temmi_hoo> they're not dissimilar :)
[04:46] <ubuntu_> Ok just one larger then the other i guess :)
[04:47] <temmi_hoo> you can build your own virtual datacenter out of virtual servers that run somewhere in them clouds
[04:47] <ubuntu_> And cloud computing is this just virtualized cluster computing
[04:47] <temmi_hoo> vagrant is one very cool tool to manage these kinds of systems
[04:49] <temmi_hoo> it allows you to build your system in very low performance model running on your laptop and then provision exactly similar cluster running in a paid cloud service such as amazon ec2 or azure cloud or googles services or somewhere else
[04:49] <ubuntu_> nice thank you so much for clearing up my confusion on those things. I still don't get the MAAS , openstack stuff yet its supposed to be for cloud /grid but it looks similar like it could be used for clustering for the new ubuntu server installs
[04:50] <temmi_hoo> vagrant can also be used to distribute your software and configuration as something called "immutable servers" so that every time you make a new version of anything, you're building a brand spanking new fresh and shiny cluster
[04:51] <temmi_hoo> now this might not be the solution most often recommended in this channel, i'm not always following what is being said in here :)
[04:51] <ubuntu_> For me when i was going thru windows 2012 r2 i got alot of the settings/how to configure stuff but  me being on such a small network  didn't get alot of the purpose so now i see the light with clustering / data center huge things
[04:51] <temmi_hoo> anyway vagrant is really cool and you can manage a system of virtual windows/linux/bsd/whatever machines with it
[04:52] <ubuntu_> like a hyper-v
[05:32] <temmi_hoo> hyper-v is the virtual machine hypervisor much like virtualbox or xen
[05:33] <temmi_hoo> vagrant is the thing that lets you control hypervisors and the virtual machines in them with programmatical configuration scripts
[09:41] <cazorla19> Does anyone know how to run service upstart job from unprivileged user withous sudo?
[09:42] <cazorla19> The issue: if I even run upstart job from unprivileged user but with sudo - the daemon process is going to be run with UID 0 which means root
[09:43] <cazorla19> It may be OK, but the issue in security: is it secure to run daemons as a root?
[09:43] <cazorla19> And how to reconfigure upstart to permit service launched as a trivial user with no sudo?
[09:44] <cazorla19> Upstart cookbook doesn't give any sense to solution: I tried to run "exec start-stop-daemon --start -u myapp --exec /usr/bin/myapp start"
[09:45] <cazorla19> But I still have upstart permission denied and also kicked out from SSH connection which looks such funny
[09:46] <cazorla19> Please if someone had this issue - help me
[13:01] <rbasak> tyhicks: I'm reviewing a merge for cpaelzer and the Ubuntu->Ubuntu diff includes this: http://paste.ubuntu.com/21398245/
[13:01] <rbasak> tyhicks: do you know if this is OK please? Does apparmor/dh-apparmor require the package to drop in these directories?
[13:01] <rbasak> I suspect it's fine, but I thought I'd check.
[13:05] <jdstrand> I can answer that
[13:05] <jdstrand> unless someone changed dh_apparmor very recently without us knowing, ntp.dirs should not have dropped those entries
[13:05] <jdstrand> tyhicks: ^ (I answered rbasak)
[13:09] <tyhicks> thanks!
[13:11] <rbasak> jdstrand: thanks!
[14:49] <ctjctj> We have a number of users that have laptops.  Those laptops are running a VM with 14.04LTS server for development purposes.  We are having an issue where the user is moving from one network to another and the VM is not picking up this change "fast enough".  What is the best way to detect that the network has changed and that we should get a new lease from the dhcp servers?
[15:09] <rbasak> ctjctj: how are the VM NICs configured? Is the host doing NAT, or are you bridging through, or something else?
[15:09] <rbasak> You could perhaps arrange to tell the VM that the cable is disconnected when the laptop is disconnected, and vice versa.
[15:15] <ctjctj> rbasak, the NICs are setup as bridged running under virtualbox for most of them (I use kvm but they don't).  So all of the VMs are true internet entities within their little lan (Or if they get IPv6, so much the better)
[15:18] <rbasak> ctjctj: my feeling is that this is best addressed at host level - by making the VM appear to have a disconnected cable at the appropriate times. I don't know to what extent virtualbox has an API-driven capability to do that though.
[15:18] <rbasak> ctjctj: together with something like ifplugd in the guest assuming you're not using network manager.
[15:20] <rbasak> ctjctj: alternatively using host-based NAT solves it too.
[15:20] <ctjctj> rbasak, *nods* I hear you and understand.  I'm not a VB expert so I don't want to go down that path.  I might just run a script every 5 minutes to do a network test to the defined router.  If it fails over 5 seconds perform an ifdown eth0; sleep 15; ifup eth0 to fetch the new configurations.
[15:21] <rbasak> ctjctj: I wonder if dhclient can be configured to attempt very frequent renewals.
[15:22] <ctjctj> That was my first hope.  I can tell it, via configuration, the data at which to renew/rebind but not a "no more than 1 hour between rebinds"
[19:42] <hackeron> Hi there, I have Ubuntu 16.04 installed on a raspberry pi, the ethernet shows up as enxb827ebd24e19. I try to create a network alias with ifconfig enxb827ebd24e19:0 192.168.88.100 but it just changes the interface IP rather than creating a second virtual interface-- any ideas?
[19:44] <compdoc> pastebin ifconfig -a
[19:45] <hackeron> compdoc: https://gist.github.com/xanview/5820c20c3500457be9754dfe7d3ae1e2
[22:59] <thenewone> hi
[23:00] <thenewone> i get this error message
[23:00] <thenewone> when i try to install something
[23:00] <thenewone> dpkg: unrecoverable fatal erro, aborting:
[23:00] <thenewone> files list file for package 'linux-headers-4.4.0-32-generic' is missing final newline
[23:00] <thenewone> E: Sub-process /usr/bin/dpkg returned an erro code (2)
[23:01] <thenewone> can how can fix this and thanks
[23:03] <RoyK> thenewone: try apt-get update again
[23:03] <thenewone> i did it 3 times
[23:03] <thenewone> with upgrade and dist-upgrade
[23:03] <thenewone> with reboot
[23:03] <sarnold> check dmesg; do you have any IO errors?
[23:04] <RoyK> and perhaps apt-get clean first
[23:05] <thenewone> i made a script with clean autoclean remove and autoremove
[23:05] <thenewone> sarnold, about that no error
[23:06] <thenewone> actualy i was trying to install docker-engine and i inturpted the installation
[23:06] <thenewone> i followed the wiki
[23:06] <thenewone> when i tryed to install it agian
[23:06] <thenewone> again*
[23:06] <thenewone> didn't work
[23:06] <thenewone> start giving this error
[23:07] <thenewone> i think i should reinstall linux-headers ?
[23:07] <van777> hi! i've got ipv6 only address now. i've made port forwarding on the router for the ip webcam. How do i access it from ipv4 address??
[23:07] <sarnold> and maybe delete the package from /var/cache/apt/....
[23:08] <RoyK> van777: nat64
[23:09] <van777> RoyK: let me google it..
[23:09] <thenewone> y
[23:09] <thenewone> sorry miss window :)
[23:15] <van777> RoyK: omg. nat64 is too hard to configure. it's built-in in cisco routers, i doubt about mine
[23:23] <sarnold> van777: you might be able to get a VPS somewhere that's dual-stacked and do some nc -l port forwarding kinds of things..
[23:25] <van777> a vps is too expensive ( i've been trying digital ocean for 4 months
[23:32] <van777> i might have luck with https://tunnelbroker.net
[23:33] <RoyK> van777: crowncloud.com is rather cheap
[23:34] <van777> RoyK: Thanks!
[23:39] <van777> Wouldn't it be easier for me to change the ISP!!
[23:41] <van777> but crownclowd cheapest plan is $4/month . not bad
[23:41] <sarnold> a tiny little instance for $15/year too. neat.
[23:42] <van777> sarnold: wow
[23:43] <van777> sarnold: heh, that offer doesn't support ipv6
[23:43] <sarnold> awwww
[23:44] <RoyK> they have ipv6 all the way
[23:44] <sarnold> RoyK: not on this one https://crowncloud.net/clients/cart.php?gid=23
[23:44] <RoyK> oh
[23:45] <RoyK> bad
[23:45] <RoyK> bad boy buddy
[23:45] <sarnold> pity, it seemed perfect :)
[23:45] <RoyK> sarnold: I have ipv6 on my vms
[23:45] <RoyK> from them
[23:46] <sarnold> yeah most of their offers do, I think that's the only one that doesn't have at least an /80
[23:46] <RoyK> I have /80
[23:46] <RoyK> anyway - it's a vm, so it's ok
[23:47] <RoyK> I just asked on #crowncloud - they are pretty good at answering there
[23:47] <RoyK> you get good support on irc from them
[23:47] <sarnold> that's worth its weight in gold :) heh
[23:48] <RoyK> the weight of those bits? :D
[23:48] <sarnold> :D
[23:49] <sarnold> RoyK: oh while you're there, "Temproary" on https://crowncloud.net/dedicated_servers.php
[23:51] <RoyK> oh - I only use their services for a mail server - I have a dedicated at bitraf.no that work well, albeit old
[23:54] <sarnold> looks cool :) but perhaps not universally useful :)