[01:24] <ko2> hi
[01:25] <ko2> Who are there?
[02:33] <pmatulis> what?
[11:55] <lordievader> Good afternoon.
[12:43] <Razva> good afternoon
[12:57] <lordievader> o/
[13:04] <shredding> Where can i learn about the stuff in ifconfig?
[13:04] <shredding> E.g. what is utun, what is ethernet etc.
[13:05] <ikonia> different device file types, you may want to look at the bridge-utils documentation
[13:07] <shredding> ikonia: Is there a place that generally gives an introduction in terms and concepts?
[13:07] <shredding> sth like https://www.digitalocean.com/community/tutorials/an-introduction-to-dns-terminology-components-and-concepts for dns
[13:07] <ikonia> shredding: no idea
[13:08] <shredding> ok
[14:56] <jayjo> How can I list all processes that are running? when I run top some PIDs come in an out of the list, but they aren't there constantly
[14:57] <Pici> ps aux
[14:57] <jayjo> for example mysqld will show up running over 3 GB but then dissapear
[14:57] <hateball> "ps auxww" shows you even more info, if that is desired
[14:57] <Pici> jayjo: you can also scroll in top, but I find htop much easier to use.
[15:04] <jayjo> ps aux and top seem to be showing me different things. top says I have 8GB memory (and I actually don't know how much I have, can I find tha from the command line... I dont have access to the AWS dashboard) but in ps aux it says the same memory is being consumed but it is 10% of memory
[15:04] <jayjo> under %MEM
[15:04] <Pici> See free -m
[15:05] <jayjo> Pici: these are MBs?
[15:05] <Pici> Yes.
[15:05] <Pici> The -m means MB
[15:05] <jayjo> ah OK, so the %MEM is not accurate I think in ps aux
[15:15] <jamespage> ddellav, coreycb: any reason where holding back on the swift 2.6.0 upload?
[15:16] <coreycb> jamespage, nope not that I know of
[15:16] <jamespage> coreycb, let me take a look - I'll sponsor it...
[15:16] <coreycb> jamespage, thanks
[15:32] <jamespage> coreycb, I'll try to unpick the python-django-compressor merge I've had on my list for a while now today as well but might be tomorrow
[15:33] <coreycb> jamespage, ok
[15:33] <coreycb> jamespage, I've started on the b3 core packages, and so far just blocked on the new paramiko release
[15:34] <jamespage> coreycb, urgh
[15:34] <coreycb> jamespage, which seems fairly important: https://review.openstack.org/#/c/268197/
[15:46] <ws2k3> im trying to install ubuntu 12.04 but at downloading installer component i just get a purple screen and then nothing happens what could this be
[15:47] <jamespage> coreycb, ~ well yes it important but generally rather than openstack specifically
[15:57] <ws2k3> im trying to install ubuntu 12.04 but at downloading installer component i just get a purple screen and then nothing happens what could this be
[15:59] <ws2k3> i also tryed other mirrors but all mirrors seems to have this issue
[16:59] <xnox> rbasak, hello =) would you like to be on the Developer Membership Board?
[16:59] <xnox> jamespage, hello =) would you like to be on the Developer Membership Board?
[16:59] <jamespage> xnox, +1 on rbasak on the DMB
[17:00] <xnox> Daviey, hello =) would you like to be on the Developer Membership Board?
[17:00] <jamespage> xnox, lol
[17:00] <xnox> jamespage, excellent. Do you have rbasak's gpg key to sign and send in nomination? =)
[17:01] <jamespage> xnox, yes
[17:01] <jamespage> coreycb, hey - looking a python-django-compressor - will need a MIR for three new packages to go to 2.0
[17:01] <jamespage> two are pull-outs of existing vendored code in 1.6
[17:02] <jamespage> and the other is new
[17:05] <coreycb> ddellav, take a look at python-pika-pool for an example of running tests as autopkgtests
[17:07] <Logos01> Howdy, folks. So I have a bit of a confusion that maybe someone here could help me work out...  why is it that there is so little interest in pulp in the .deb world?
[17:07] <coreycb> jamespage, need a hand with those MIRs?
[17:07] <Logos01> I mean, it's been over two years since the basic framework for .deb support was published to the pulp github, and it's *still* nonfunctional.
[17:07] <jamespage> coreycb, I'll need a ffe first - you focus on the rest of the stuff...
[17:07] <coreycb> jamespage, ok
[17:11] <rbasak> What's pulp?
[17:11] <Logos01> Pulp is a repository synchronization/publication engine
[17:11] <rbasak> Why does the .deb world need this?
[17:11] <Logos01> http://www.pulpproject.org/
[17:12] <Logos01> rbasak: Well for one it would enable the use of products like Katello
[17:12] <rbasak> What's Katello? Why does the .deb world need that?
[17:13] <Logos01> It is a significant gain over apt-mirror in that it enables administrators to create repositories interanally and associate them with specific servers in order to more readily manage what packages are accessible to what servers (up to and including recency of mirror, to enable package lifecycles for promotion paths from nonprod to prod, etc., etc..)
[17:14] <Logos01> Katello is the opensource upstream of the Red Hat Network Satellite (version 6.x)  ( http://www.katello.org/ )
[17:14] <Logos01> It ... does a great deal of things, including bundling foreman, puppet, pulp, and candlepin
[17:14] <Logos01> Gives you recency, inventorying, ownership tracking, config management, etc., in a bundled manner.
[17:15] <rbasak> I don't see a benefit here. You need to frame this in terms of the benefits it brings, not in terms of other things whose benefit you have also not explained.
[17:15] <Logos01> O_o
[17:16] <rbasak> IM(personal)HO, apt repositories outside the distro are fundamentally broken anyway. PPAs are about as far as they'll stretch.
[17:16] <rbasak> apt wasn't designed for this, and it shows.
[17:16] <rbasak> I suspect RPMs are the same.
[17:16] <Logos01> rbasak: You can do PPAs as well with this, but it's primary function is in handling the core distro channels.
[17:16] <Logos01> It's a repository mirroring engine.
[17:17] <jamespage> coreycb, for reference:
[17:17] <jamespage> https://bugs.launchpad.net/ubuntu/+source/python-django-compressor/+bug/1554134
[17:17] <Logos01> It lets you do things like expose the current packageset to your dev environment, a week old version of that to your acceptance/staging/testing environments, and a month old to your production
[17:18] <rbasak> That sounds handy.
[17:18] <Logos01> So you can just let packages autoupgrade and still have that assurance of burn-in on all packages.
[17:18] <Logos01> This is what I mean by "recency"
[17:18] <rbasak> Make it do that easily for Ubuntu users and perhaps they'll take it up.
[17:18] <Logos01> It *ALSO* lets you do things like track which DSAs (or the Ubuntu equivalent) are applicable to which of your servers.
[17:18] <Logos01> rbasak: ... That's exactly what I'm asking about.
[17:18] <rbasak> Add a million layers of abstraction in the middle, and I suspect that nobody will bother.
[17:19] <Logos01> Why isn't this there already? It just needs plugin development.
[17:19] <Logos01> Apparently the answer is nobody in the .deb world has heard of it. Which seems surreal to me.
[17:19] <rbasak> Someone needs to champion it for Ubuntu I guess.
[17:19] <Logos01> You're familiar with Ubuntu Landscape?
[17:19] <rbasak> Yes.
[17:19] <coreycb> jamespage, thanks
[17:19] <Logos01> Katello is a free and opensource "competitor" in a sense to it.
[17:20] <Logos01> (You can also use pulp for pip, cpan, gems, or puppetforge)
[17:20] <rbasak> So be competitive :)
[17:20] <Logos01> I'm not a developer.
[17:20] <rbasak> Ah.
[17:20] <Logos01> I'm just trying to understand why it seems like nobody in the .deb world seems to give a crap about it.
[17:20] <rbasak> What you're really asking is why nobody is spending money on developing it.
[17:20] <Logos01> I've been asking this for years. I never get a good answer beyond "I've never heard of it"
[17:20] <Logos01> Which is just ... bizarre.
[17:21] <rbasak> There's a big difference between those questions.
[17:21] <rbasak> Nobody has heard of it because nobody else is spending resources on developing it for this world.
[17:21] <rbasak> Simple answer I think.
[17:22]  * Logos01 shrugs
[17:22] <Logos01> It's not even like it would take a great deal of manhours to do.
[17:22] <Logos01> https://github.com/pulp/pulp_deb
[17:22] <rbasak> Give somebody a good reason to spend those manhours, and perhaps they will.
[17:24] <Logos01> That is ... a surreal answer, to me.
[17:24] <rbasak> That is exactly how the Free Software ecosystem works. I don't understand why it's surreal.
[17:25] <Logos01> Because the benefits of the thing are rather extensive; knowledge of the project in and of itself should comprise sufficient reason for at least *someone* to want to make that development.
[17:25] <Logos01> Especially considering it is developing a plugin, not overhauling the engine itself, that is all that's called for.
[17:25] <rbasak> That's not how it works.
[17:25] <Logos01> And yet no matter how many times I've brought this up over the years nobody has ever seemed to ever say anything other than "Why should we care?"
[17:26] <Logos01> And I can explain and reexplain the benefits of having repository management and all the ancillary benefits therein and it just doesn't sink in.
[17:26] <rbasak> You can wish for developers to magically appear all you want. Actually everyone only does stuff because of some reason to motivate them. This applies to everything, not just this project.
[17:28] <Logos01> There's a difference between wishing for something to have happened, and trying to understand why there seems to be significant resistance to something that has obvious and imminent gains for the community.
[17:28] <Logos01> The latter is what I'm trying to figure out here.
[17:28] <rbasak> I don't see any resistance.
[17:28] <rbasak> You say resistance, I say lack of motivation.
[17:29] <Logos01> That is; not merely, "Why hasn't someone done what *I* want" but rather "Why has no one done this despite how painfully obvious the benefits of it are?"
[17:29] <Logos01> And to be clear; they *are*
[17:29] <Logos01> Which is why I find answers like yours surreal.
[17:29] <Logos01> Because this is far from the first time I've brough this topic up.
[17:29] <rbasak> I suspect most people in a position to need this find it easier to just pay for Landscape.
[17:29] <Logos01> Yet that is *always* the answer: "Why should we care?" "Do it yourself", etc..
[17:30] <rbasak> s/need/develop/ maybe.
[17:30] <rbasak> Again, that's how the ecosystem works.
[17:30] <Logos01> Or, more commonly, as you did, "Never heard of it"
[17:30] <Logos01> And it's that last part that is the most perplexing.
[17:31] <Logos01> Oh well.
[17:31] <Logos01> It seems the answer hasn't changed from the last dozen times I've asked this question (in here, no less)
[17:31] <Logos01> Thanks for your time.
[17:31]  * Logos01 walks away in utter astonishment
[17:31] <rbasak> That's because the motivation for the existence of the Free Software ecosystem hasn't changed, either.
[17:32] <rbasak> How can you still be astonished, if others have given you the same answer previously?
[17:32] <nacc> well, it seems like pulp explicitly lists only RH-family distros, no? and the diagram is all yum ... http://www.pulpproject.org/
[17:32] <jamespage> matsubara, rbasak: did you guys make progress on the dlm merge from debian?
[17:32] <nacc> or is that the point of this discussion?
[17:33] <rbasak> jamespage: I don't recall any movement on that.
[17:33] <rbasak> jgrimm: do you know the status of the dlm merge?
[17:33] <nacc> "There is also a community-contributed plugin for Debian packages."
[17:33] <nacc> Logos01: --^ ?
[17:33] <nacc> Logos01: why isn't that "good enough"?
[17:34] <Logos01> nacc: It's a skeleton that is non-functional and has been in that state since it was first created over two years ago.
[17:34] <Logos01> It wouldn't take much work to finalize but nobody in the .deb world ever seems to care.
[17:35] <rbasak> It sounds to me that it's a RH world product that claims to be "cross platform" in order to make it more appealing to users to lock themselves in, because in reality they're only paying lip service to that claimed portability.
[17:35] <nacc> Logos01: i have to agree with rbasak; what you've described above is a generic argument for why pulp is good. But not why I should care to contribute. And tbh, it doesn't solve a gap for me, as a user. The gap seems to be deploying pulp in a place that wants to host Debian
[17:35] <nacc> or Ubuntu
[17:36] <Logos01> nacc: I'm not agitating for it to be done.
[17:36] <Logos01> I'm not *asking* for it to be done.
[17:36] <Logos01> I'm trying to understand why nobody in the .deb world already *wants* it to be done.
[17:38] <nacc> Logos01: why would anyone in the .deb world particularly care to make a RH/rpm-based product integrate with deb-based? why is it so much better than what already exists? (and if it is, i think it would naturally be supplanting it in the ecosystem ... the best tools, ime, exist & thrive in FOSS)
[17:39] <jamespage> rbasak, I'll deal with it - clearing my merge backlog...
[17:39] <Logos01> nacc: The closest analogue to it is apt-mirror, and pulp is a significant gain over that in terms of the raw functionality it provides, to the point where I don't even know where to begin to describe how much so this is true.
[17:40] <Logos01> And you're asking exactly the question I'm asking with the latter parts there. Which is why I am constantly amazed by the lack of penetration or care/concern it has here.
[17:40] <rbasak> The people who might want that functionality probably use Landscape. So you need to be comparing against Landscape, not apt-mirror.
[17:40] <Logos01> rbasak: Pulp is best compared against apt-mirror.
[17:41] <Logos01> Katello, which uses pulp as a component, is what is best compared against Landscape.
[17:41] <rbasak> If you say so. I doubt users care.
[17:41] <Logos01> You can use Pulp without anything else. Many do.
[17:42] <Logos01> Also, and this is a major consideration point; Individuals who want to learn how to manage the closed-source product from Red Hat (RHN Satellite) can deploy their own Katello instances in the lab, and take exactly those skills across without having to license anything.
[17:42] <Logos01> Makes for a much easier entry path into competence.
[17:44] <rattking> I use aptly snapshots to do what sounds like the same things
[17:44] <rattking> same things being alpha/beta/prod snapshots on a local mirror plus personal repos
[17:47] <Logos01> rattking: From the description (on the page as well) it seems very similar, yes.
[18:55] <nacc> jamespage: i just noticed something else about commons-vfs that I missed before, I apologize! there's a bindep from libcommons-vfs-java-doc -> libcommons-net-java-doc but the latter is in universe. It looks like historically (it hasn't been packaged in ubuntu since oneiric?) it was in main. Should it be MIR'd? the package does build & test successfully without either of libjackrabbit or
[18:55] <nacc> libcommons-net-java-doc, but i'm guessing maybe some documentation linking won't work?
[19:45] <jancoow> Hi. Does someone know a tutorial (that works) for running at least 1 application over a vpn, but the rest of the server over the normal ethernet connection
[20:17] <lordievader> jancoow: VPN's usually work on a basis of routes, not applications.
[20:24] <bekks> jancoow: lordievader: https://schnouki.net//schnouki.net/posts/2014/12/12/openvpn-for-a-single-application-on-linux/
[20:25] <lordievader> bekks: 'That page doesn't exist!'?
[20:26] <lordievader> Ah, double schnouki.net
[20:26] <bekks> indeed - no clue how I copied that :D
[20:27] <lordievader> This forces a program to use a different default gateway?
[20:27] <bekks> Yes. Which basically what a VPN does, as well.
[20:27] <jancoow> well i tried that
[20:28] <jancoow> the virtual interface is kinda working but it lost connection after x seconds
[20:28] <jancoow> like i could ping , but when i try 10 seconds later: nope no connection
[20:28] <lordievader> Nice though.
[20:28] <jancoow> (and that's without vpn running)
[20:28] <lordievader> jancoow: So, the underlying network connection has problems?
[20:29] <bekks> jancoow: that solution requires a vpn.
[20:31] <jancoow> bekks: i also tried a vpn. If i do it in "the first 10 seconds" the vpn could communicate with the vpn server. But this also timed out after 15 seconds or so
[20:31] <jancoow> lordievader: i do think so!
[20:32] <jancoow> i simply runned this https://gist.github.com/Schnouki/fd171bcb2d8c556e8fdf with ""up"
[20:32] <jancoow> and changed the fping to ping (because i doesn't have fping installed) :)
[20:33] <jancoow> and eventually started openvpn (with another config, ofcourse)
[20:33] <lordievader> jancoow: So fix your network first ;)
[20:33] <jancoow> lordievader: well the normal ethernet connection is just fine (the one on eth0)
[20:34] <lordievader> < jancoow> lordievader: i do think so! <- /me is confused
[20:34] <jancoow> wut ;p
[20:35] <lordievader> I asked if the underlaying network connection has problems, you said yes. And later you said it didn't.
[20:36] <jancoow> lordievader: oh wel. The "Yes" was more on the underlaying network connections UNDER the vpn :). So actually the bridge between the virtual interface and the realinterface
[20:36] <lordievader> Ah, the regular network is fine?
[20:37] <jancoow> yes!
[20:37] <jancoow> server is already up for 200 days and never had problems :)
[20:39] <lordievader> Ah, I misunderstood. So after some time the vpn connection, and only the vpn connection, dies?
[20:40] <jancoow> not only the vpn connection. Even when i doesn't start the vpn on it, and i'm trying to use ping it doesnt work
[20:41] <lordievader> jancoow: Was the "ip netns exec frootvpn ping www.google.com" from the guide set?
[20:41] <jancoow> yes
[20:41] <jancoow> this is for example the output after i imdetialy run the script: https://jancokock.me/f/b566a
[20:42] <lordievader> Does ping work after removing that setting?
[20:42] <jancoow> how can i test that without running that command?
[20:43] <bekks> You cant. :)
[20:43] <lordievader> You cant remove those settings?
[20:44] <jancoow> well it isn't a setting, right? It exec's a program, and that programm is ping
[20:44] <bekks> No. The exec program is frootvpn
[20:44] <bekks> and that executes ping.
[20:44] <jancoow> ah!
[20:45]  * lordievader fears his understanding of the Linux network stack is too little
[20:45] <bekks> actually that frootvpn is a network namespace which you execute ping in.
[20:46] <bekks> So it is vital to use "ip netns frootvpn ..."
[20:46] <jancoow> but why does it stops working after some seconds :/
[20:46] <jancoow> that's the question ;
[20:46] <jancoow> ;p
[20:47] <jancoow> do you know if there are somewhere logs ??
[20:47] <bekks> Look at the logs? :)
[20:47] <jancoow> ?*
[20:47] <bekks> Logs are in /var/log/ e.g.
[20:47] <jancoow> yeah
[20:48] <jancoow> there are a lot of logs there :D
[20:48] <bekks> Thats why the directory is names /var/log :P
[20:48] <jancoow> the only thing i can see in dmesg is [16873278.337995] IPv6: ADDRCONF(NETDEV_UP): vpn0: link is not ready
[20:48] <jancoow> [16873278.360117] IPv6: ADDRCONF(NETDEV_CHANGE): vpn0: link becomes ready
[21:03] <jancoow> bekks: any idea? :D
[21:29] <jancoow> btw
[21:29] <jancoow> is it bad that i didn't rebooted my server for already 200 days
[21:30] <genii> If you did security updates in that time and didn't reboot yet, then yes
[21:30] <jancoow> just regual updates in apt update/upgrade
[21:31] <tomreyn> during the past 200 days there was the glic patch, which suggests you should reboot
[21:33] <tomreyn> (not all ubuntu versions / platforms were affected by it, though). while i have not checked, it is somewhat likely that kernel security patches were also made available during this time. which, unless you use live kerbel patching, would also requiore a reboot to apply.
[21:33] <tomreyn> s/ kerbal / kernel /
[21:33] <jancoow> ubuntu doesn't have the ability for live kernal patching right?
[21:34] <jancoow> only redhat right?
[21:34] <tomreyn> linux has the ability
[21:34] <tomreyn> in some versions or other
[21:37] <genii> Ubuntu does, but it gets convoluted
[21:37] <genii> !info ksplice
[21:38] <jancoow> and how safe is it to do?
[21:38] <jancoow> or is it stable
[21:39] <genii> ksplice itself is stable, but the updates then work differently than normal
[22:01] <coreycb> ddellav, can you update the (build-)dependencies for heat?
[22:01] <coreycb> ddellav, looks ok other than that
[22:01] <ddellav> coreycb ok
[22:06] <Logos01> jancoow: If you're worried about live kernel patching then you might want to look into LXD
[22:06] <Logos01> And then containerize all of your services.
[22:06] <Logos01> Probably a safer/easier path
[22:35] <jancoow> i'm not woried
[22:35] <jancoow> just asking :)
[22:35] <jancoow> still didn't fix my vpn
[22:50] <ChibaPet> Hey all. If this is not yet on-topic I'll go elsewhere, but Xenial installs with the server ISO that set up LUKS root seem to all time out on shutdown, as systemd can't dispense with the LUKS container or the MD-RAID (in some cases) underneath it. Is there a known workaround
[22:50] <ChibaPet> ?
[22:59] <Jukse> Hello! i have a question!
[23:22] <ChibaPet> !ask
[23:25] <sarnold> ChibaPet: probably best to file a bug against systemd, sounds like something that might require attention from someone in foundations
[23:26] <ChibaPet> sarnold: Yeah, I suspect that's the thing to do. I'll also read a bit about this shutdown-initrd concept.
[23:26] <ChibaPet> I wish I'd noticed this sooner, as we're somewhat close to release.
[23:27] <ChibaPet> Anyway, commuting, and I'll file a bug report when I am connected via wires again.
[23:27] <sarnold> yeah, I'm surprised just how soon we are..
[23:27] <sarnold> mm wires