[01:06] <JRWR> Trying to setup GlusterFS {3.5.2}, Both boxes are connected over a tinc connect and confirmed working ok, glusterfs is having a hard time adding peers, it just hard locks and never completes, it just locks up with the log line of 0-glusterfs: transport.address-family not specified. Could not guess default value from (remote-host:(null) -- my command is gluster peer probe 10.10.xx.xx
[03:52] <hallyn> hm, rabbitmq-server wont' install for me in xenial, either a host or ina container. systemd job fails.
[03:56] <hallyn> odd, no new upload since january
[03:59] <hallyn> hm, no, in a container it starts fine with /etc/hosts fixed up.
[08:29] <BusyElf> Ubuntu 14.10 Server : I have a problem with a programme shutting down at 02:30 -ish every morning. Nothing in the log files to sugges it was a system side order to shut down.. does anyone have any ideas?
[08:31] <BusyElf> Hi :-)
[08:33] <BusyElf> Ubuntu 14.10 Server : I have a problem with a programme shutting down at 02:30 -ish every morning. Nothing in the log files to sugges it was a system side order to shut down.. does anyone have any ideas?
[08:33] <TJ-> !eol | BusyElf 14.10 is no longer supported
[08:34]  * BusyElf facepalms
[08:35] <TJ-> BusyElf: that's not going to cause your issue, but it sounds like an application-specific thing, unless you've a system cron job causing it
[08:36] <BusyElf> we did have a cronjob in place to close it and restart it after backup. Now there's no backup we deleted the cronjob  but it still shuts down
[08:36] <BusyElf> The server has since been restarted as well
[08:39] <BusyElf> sorry My mistake, it's Ubuntu 14.04.4 LTS version
[09:01] <halvors> Anyone knows why mod_http2.so is missing from the apache2 package in Ubuntu Server 16.04?
[09:01] <halvors> I have the latest build.
[09:01] <halvors> The http2.load and http2.conf is there. Just not the shared object.
[09:05] <lukasa> halvors: The Ubuntu developers elected not to add mod_http2 to the LTS release packages
[09:13] <halvors> Well, they did anyway.
[09:15] <halvors> lukasa: It is present in the latest package.
[09:15] <halvors> Just not the so file.
[09:15] <halvors> So it seems to be there for the user.
[09:18] <TJ-> halvors: http2 for both apache2 and nginx, as I recall, are deliberately not included due to the 'experimental' nature
[09:20] <TJ-> halvors: as the package changelog says: "- Don't build experimental http2 module for LTS:"
[09:21] <TJ-> halvors: oh, seems like it has been enabled in nginx, as of start of April. The original discussion was that neither package would have http2
[09:56] <TJ-> who handles building the cloud-images for the cloud-images and partner-images archives?
[10:16] <halvors> TJ-: There is apache2 config present for this: /etc/apache2/mods-available/http2.load
[10:16] <halvors> Just that the .so file is missing.
[10:16] <halvors> It should be if it's going to be disabled...
[10:17] <TJ-> halvors: I agree; have you reported the bug?
[11:42] <frickler> jamespage: yet another issue involving systemd task accounting https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1567381 , pretty significant for our deployment
[11:44] <jamespage> cpaelzer, rharper: ^^ thats probably quite an important bug to resolve
[11:44] <jamespage> thanks frickler - your pre-release feedback has been invaluable!
[11:46] <frickler> getting some feedback on https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1564812 would also be nice, not sure whether you create the sudo config or that is taken from upstream
[11:47] <cpaelzer> jamespage: yeah that looks bad from just the description
[11:48] <cpaelzer> jamespage: since it is about libvirt also highlighting hallyn and smb to bug 1567381
[12:03] <frickler> jamespage: well, thanks for being so supportive, much better than having to go through this bug-fixing for another 6 months after the release ;)
[12:14] <ddellav> jamespage manila is ready for review and push: lp:~ddellav/ubuntu/+source/manila i rebased against ubuntu-server-dev
[12:19] <smb> jamespage, hm, practically the qemu processes are part of the virtualisation
[12:20] <smb> so from my thinking they should be accounted against it. Maybe the question is to adapt the task limit
[12:22] <frickler> smb: I would rather vote that this is similar to sshd, there the processes of users logged in via ssh also should not be counted towards the ssh.service
[12:25] <smb> frickler, but there VMs started by libvirt directly relate to the service. processes started by users logged in is one level deeper
[12:27] <frickler> smb: hmm, maybe, but then the question is how to set a useful limit. I get 490 of of 512 tasks being used by starting 4 cirros instances, the 5th already fails
[12:29] <smb> frickler, Yeah, probably a question what this really is accounting for. I suspect that is might rather be threads than tasks
[12:31] <frickler> smb: "tasks" for systemd means threads, yes, the process count is 7 ( 4 x qemu, 2 x dnsmasq, 1 x libvirtd)
[12:44] <smb> and the number for qemu then could vary a lot depending on the config of the VM (possibly one per disk/nic though I have not verified that). Not sure how to fit that best into the accounting. Either the limit must be rather high or ... hm I assume this is all depending on cgroups, so libvirt would need to place each qemu/VM into an independent one. Which might be a rather big change.
[13:16] <patdk-wk> sarnold, dunno what happened, but whatever was messed up with grub efi boot + luks, got fixed in beta
[13:59] <rsevero> Hi. One of my Ubuntu 15.10 servers started this morning to present errors like: "Setting up thermald (1.4.3-5ubuntu1) ...
[13:59] <rsevero> Installing new version of config file /etc/thermald/thermal-conf.xml ...
[13:59] <rsevero> Error getting authority: Error initializing authority: Error calling StartServiceByName for org.freedesktop.PolicyKit1: Timeout was reached (g-io-error-quark, 24)
[13:59] <rsevero> Failed to execute operation: Connection timed out
[13:59] <rsevero> Error getting authority: Error initializing authority: Error calling StartServiceByName for org.freedesktop.PolicyKit1: GDBus.Error:org.freedesktop.DBus.Error.TimedOut: Failed to activate service 'org.freedesktop.PolicyKit1': timed out (g-dbus-error-quark, 20)
[13:59] <rsevero> " during an "apt-get dist-upgrade"
[14:00] <rsevero> It presented errors like the above during a boot I did just before the upgrade. Any idea what might be wrong? And how to fix?
[14:24] <coreycb> ddellav, I pushed your barbican changes, and we can pick those up when we do the final release of barbican
[14:29] <ddellav> coreycb: ok thanks
[14:36] <jamespage> ddellav, doing manila now...
[14:37] <jamespage> coreycb, do you want me to sweep up the 2.0.0 as well for manila whilst I'm doing ddellav's manila-share landing?
[14:40] <coreycb> jamespage, sure!
[14:41] <jamespage> coreycb, did all of the releases already come out?
[14:41] <jamespage> 07:47 this morning?
[14:42] <coreycb> jamespage, I believe so, that's what openstack-announce says
[14:44] <jamespage> coreycb, okies...
[14:50] <coreycb> jamespage, ugh, the new keystoneauth1 adds a new dep for python-betamax
[14:51] <jamespage> can we pick a compromise version?
[14:51] <coreycb> jamespage, we can stay where we are or I can patch out the tests, it's just for tests
[14:51] <jamespage> coreycb, i'd prob go with patching out the tests..
[14:51] <coreycb> jamespage, ok
[14:51] <jamespage> coreycb, does it fix important things?
[14:52] <coreycb> jamespage, there are some bug fixes, 2.4.0 gets us to the  upper constraint
[14:53] <coreycb> I'll go with patching the test out for now, that's what upstream's been using
[14:54] <jamespage> coreycb, okies manila uploaded
[14:54] <coreycb> jamespage, nice
[14:55] <ruben23>  hi guys i have installed apache2 on ubuntu server, when i run my site i get this -----> 404 NOt found any idea guys how to troubleshoot this..?
[14:58] <jvwjgames> Apache error logs
[14:58] <jvwjgames> Logs are your best friend
[15:01] <ruben23> http://pastebin.com/M5DzF2J3  <--------- nothing significant to error appearing
[15:14] <teward> ruben23: 404 would suggest that what you're trying to access doesn't exist
[15:14] <teward> ruben23: I'd set a more verbose error logging level and try again
[15:14] <teward> or, nuke your browser cache first
[15:14] <teward> then try again
[15:16] <_max_> hi guys, quick question: does anybody here have experience with fai-project under ubuntu?
[15:25] <BlessJah> I'm trying to setup PXE boot using iso as local repository. Debian-installer tries to contact my repo (/cdrom/dists exposed via http) and succeds to find dists/trusty, bet then tries trusty-updates which is not available on cd and hangs when it receives 404. (I've figured out -server channel is better place to ask such question)
[15:36] <RoyK> BlessJah: mirror the whole repo, not just what's on the ccd
[15:36] <RoyK> BlessJah: mirror the whole repo, not just what's on the cd
[15:37] <BlessJah> RoyK: how large whole repo could be, and why is cdrom not enough?
[15:38] <BlessJah> I'm actually trying to bootstrap local pxe/repo mirror server for environment behind firewall
[15:38] <RoyK> got a proxy?
[15:39] <BlessJah> not at the moment of bootstrapping
[15:40] <RoyK> I don't know how large those repos are - I guess 10-20 gigs
[15:40] <RoyK> https://help.ubuntu.com/community/Rsyncmirror
[15:40] <RoyK> see that
[15:40] <BlessJah> way too big
[15:41] <RoyK> why?
[15:41] <Karunamon> so here's a question that probably doesn't come up often - anyone familiar with setting up serial terminals?
[15:41] <RoyK> BlessJah: that's not really a lot
[15:41]  * RoyK setup a neat little server with 80TiB disk space yesterday
[15:41] <RoyK> Karunamon: I am
[15:42] <Karunamon> RoyK: I've got a wy50 hooked up to my machine and giving me a login session, but it's still really easy to screw it up so that a reset is required to become usable again
[15:42] <BlessJah> RoyK: because it has to fit usb drive, 8GiB is max, >1GiB is good
[15:43] <Karunamon> i'm almost certain that it's a charset thing.. I can use sh, but not bash, opening anything like mutt or vim completely trashes it
[15:43] <RoyK> BlessJah: I mean on the repo, not on the final install
[15:44] <Karunamon> that or some disagreement between my tty settings and the settings on the terminal
[15:44] <RoyK> Karunamon: ah - not sure, sorry. try setting TERM=vt100
[15:44] <RoyK> Karunamon: as in "export TERM=vt100
[15:44] <RoyK> "
[15:45] <jamespage> coreycb, I have the theme refresh from design - lemme take horizon
[15:45] <coreycb> jamespage, awesome
[15:45] <Karunamon> nope, that doesn't appear to work
[15:45] <BlessJah> I know that it's the repo that is 20GiB, but all files required for bootstrap has to fit pendrive so they'd fir pendrive and be manageable to be sent over crappy internet
[15:45] <RoyK> BlessJah: then I don't know
[15:46] <_max_>  i try to build a ubunut nfsroot from/with "trusty archive.ubuntu.com/ubuntu"  and i get a "not configured" error for udev, plymouth and initramfs-tools depending on udev, but if i manually try to install udev on 14.04 i have no problems
[15:47] <_max_> any ideas why?
[15:51] <BlessJah> I know my limits may sound strange, but they
[15:52] <BlessJah> 're pretty reasonable: to be able to perform PXE boot using just iso image
[16:36] <dr4c4n_> hi, I'm having trouble setting a static ip address on a box to create a management network that's running kvm
[16:37] <dr4c4n_> do I need to create a bridge for that?
[16:39] <sdeziel> dr4c4n_: what trouble are you having when setting the static IP?
[16:40] <dr4c4n_> sdeziel: what I'm trying to do for now, is set a static ip on the kvm host machine, and a static ip on the one machine I want to use to manage it, but I'm getting destination host unreachable, both are connected to a switch, and ip link show, shows that I'm using the right interfaces on each
[16:41] <dr4c4n_> and ifconfig shows that both machines have their ip addresses assigned
[16:42] <sdeziel> dr4c4n_: try to start a ping stream from you host and use tcpdump on the destination to confirm your traffic reaches it
[16:47] <coreycb> jamespage, keystonemiddleware 4.4.0-3 is in the xenial queue.  I've asked inifinity if he can accept that before any core packages since it is needed at build time to generate the right configs.
[16:49] <coreycb> jamespage, he just accepted middleware
[16:49] <jamespage> coreycb, lol following along
[17:06] <jvwjgames> I have an issue with curl
[17:07] <jvwjgames> I am trying to use my server to try to send push notifications and it keeps on saying error
[17:10] <ratrace> like what
[17:10] <crazybluek> can't remember  why Iv
[17:10] <jvwjgames> Are you talking to me
[17:11] <ratrace> jvwjgames: yup. error like what?
[17:11] <crazybluek> I've set up mine ubuntu server with localhost 127.0.0.1 and pc_box_name 127.0.1.1
[17:12] <jvwjgames> http://picpaste.com/pics/IMAG0456-58PGayJB.1460049132.jpg
[17:13] <sarnold> crazybluek: that's normal; 127.x.x.x. goes via loopback and never leaves the machine
[17:14] <ratrace> jvwjgames: -d flag sets the post body string. you're using it wrong, it should be an "urlencoded" string
[17:14] <sarnold> jvwjgames: easy, call google and ask them to give you an API key for their gcm-http service
[17:15] <crazybluek> sarnold ok
[17:15] <sarnold> crazybluek: some applications get cranky if they forward-resolve an address then reverse-resolve the address and get something different
[17:15] <sarnold> crazybluek: this configuration lets the computer be used for one set of addresses, localhost for another set, and the two won't stomp on each other
[17:16] <crazybluek> ok
[17:16] <crazybluek> trying to figure out if there are some misconfiguration on  pc/ubuntu
[17:17] <ratrace> jvwjgames: -d "to=/topics/foo-bar" -d "data=message"        or as single string -d "to=/topics/foo-bar&data=message"  however, as I said, it must be urlencoded.
[17:17] <jvwjgames> Ok
[17:19] <jvwjgames> Maybe this will help
[17:20] <jvwjgames> Http://goo.gl/w7Yrg7
[17:21] <ratrace> jvwjgames: help what?
[17:22] <jvwjgames> Cause I am trying to URL encode that code
[17:22] <jvwjgames> Where it says http post request
[17:23] <ratrace> jvwjgames: for starters, curl is not really a good tool for structured data. But if you must, I suppose using application/json content type is easier because you don't have to urlencode.
[17:24] <jvwjgames> Then how do I use it then
[17:24] <ratrace> use a programming language that understands structured data like perl or python.
[17:24] <jvwjgames> Ok
[17:25] <jvwjgames> Thanks
[17:25] <jvwjgames> So just slap the code into a .py file and execute it
[17:26] <jvwjgames> Right
[17:26] <ratrace> and if you wish to stick to curl, you can try the "HTTP POST Request" example from that second link, using --header "Content-Type: application/json" and -X POST and  -d '{"to": "/topics/foo-bar", "data" ....   '
[17:26] <ratrace> put that whole json structure into -d string
[17:26] <ratrace> you can also put it in a file and use -d @filename for curl
[17:28] <ratrace> jvwjgames: there's a python client for GCM so you don't have to reinvent the wheel: https://github.com/geeknam/python-gcm
[17:28] <ratrace> surely a programming language of your choice might have a lib/binding for that API
[17:50] <jvwjgames> ratrace thanks for your help
[17:51] <jvwjgames> Returned http 200 Ok
[17:51] <jvwjgames> And message arrives at my phone
[17:53] <ratrace> jvwjgames: yw :)
[17:56] <jvwjgames> I ended up adding the json code in a .py and just having curl
[17:56] <jvwjgames> https://gcm-http.googleapis.com/gcm/send @GCM-sender.py
[17:56] <jvwjgames> And that worked
[17:57] <jvwjgames> *-d @GCM-sender.py*
[18:00] <jvwjgames> And now to automate that with cron jobs witch I can do
[18:00] <jvwjgames> Again thanks for your help
[18:00] <ratrace> jvwjgames: you shouldn't. even though extensions don't define mimetype, you shouldn't use .py if it's not python code, and it isn't, it's JSON.
[18:01] <wsirccc> Wer hat Lust auf nen Smalltalk über Paketbau mit mir.
[18:01] <jvwjgames> So even though it works it's not good practice
[18:01] <ratrace> jvwjgames: exactly.
[18:02] <jvwjgames> So use .sh
[18:02] <wsirccc> pardon: how likes smalltalk about packaging in broader sense with me?
[18:02] <sarnold> wsirccc: "who" :)
[18:04] <wsirccc> you?
[18:04] <wsirccc> (;
[18:05] <sarnold> wsirccc: it depends on the specific question; I only do simple things..
[18:06] <wsirccc> sarnold: Are you interested?
[18:06] <sarnold> sure
[18:09] <wsirccc> Well in my case we speak of packaging a program, which must be developed first, by and by.
[18:09] <wsirccc> Got me?
[18:10] <wsirccc> it is about runnable qemu image on finger snip.
[18:11] <sarnold> "finger snip"?
[18:12] <wsirccc> at the finger tips
[18:13] <sarnold> aha :) so you want to package something like a qcow2 image and a command line that launches the VM?
[18:13] <wsirccc> yes
[18:13] <wsirccc> thats part of it.
[18:14] <wsirccc> And now I ask the packager for experience in user program specifics.
[18:14] <wsirccc> That describes the situation.
[18:15] <wsirccc> launch? I want it to be "fully scriptable".
[18:16] <wsirccc> controllable over ssh out of box as a user interface
[18:16] <wsirccc> So what does it take?
[18:16] <wsirccc> Boss?
[18:16] <wsirccc> (;
[18:17] <wsirccc> sarnold
[18:17] <wsirccc> sarnold:?
[18:19] <wsirccc> Is it possible and how much time one should invest for it?
[18:20] <sarnold> wsirccc: it depends how far you want to go with it. you could do something similar via lxd quite quickly, but containers aren't VMs. if you -want- it to be qemu, then you've got a lot more work.
[18:22] <wsirccc> "got a lot more work" about which steps works you speaking
[18:22] <wsirccc> ?
[18:22] <nacc> and by "do something via lxd", sarnold means lxd already does this
[18:23] <sarnold> wsirccc: lxd is easy, you pull down an image, configure it as you need, then publish the modified image
[18:24] <zzxc> Hey I have an apache issue. I have multiple site-enabled. Currently I have a conf for each of foo.com/a foo.com/b foo.com/c. I'm wanting to handle multiple another domain url for bar.com/a. So I went with virtual_host *:443 and servername foo.bar with a location /a in one conf, /b in another /c in a third,
[18:25] <zzxc> then another file that has the same virtual host with the servername of bar.com. bar works withouth any issue, but only the first of the foo.bar will load. All foo.com confs are disregarded.
[18:26] <zzxc> unload adv_windowlist.pl
[18:26] <zzxc> err sorry about that.
[18:30] <wsirccc> sarnold
[18:30] <wsirccc> sarnold
[18:30] <wsirccc> sarnold: thanks, you send me to another project. Understand from it that it is another tool like uvtools that download images from the test server, patch them start them and controll them by ssh.
[18:31] <wsirccc> Is it that perfect?
[18:31] <sarnold> wsirccc: I'm sorry to say that I never got the hang of uvtools
[18:32] <sarnold> wsirccc: uvtools uses simplestreams to distribute information about images and that's very nearly a black-box -- it's impossible to learn more about it, it's not very well documented, and the tools that exist have no options for debugging..
[18:33] <wsirccc> Well, I am definitely searching something more: my the automation of physical machines shall be test es well.
[18:34] <wsirccc> Just recall uvtools because soone else mentioned it before here in the chat.
[18:35] <wsirccc> sarnold: "got a lot more work", which one ?
[18:35] <sarnold> you certainly could distribute raw images and virsh command line scripts to import the VM but assuming people have a working libvirt install may be asking a lot too
[18:37] <wsirccc> sarnold: Do you know anybody, who does it?
[18:38] <sarnold> nothing that does everything you want
[18:38] <wsirccc> I could distribute the creating, preseeding script es well, as templates, voala
[18:40] <wsirccc> sarnold: can one asume that there might be really noone how offers ready to use that qcow2 machines?
[18:41] <sarnold> wsirccc: you could either package a shell script that starts kvm with all the needed parameters
[18:41] <sarnold> wsirccc: or you could package a shell script that calls virsh with the needed parameters
[18:41] <wsirccc> why not qemu?
[18:41] <sarnold> wsirccc: but different users will have different networking needs, differentinput devices they want passed through, etc
[18:42] <sarnold> kvm/qemu, same thing :)
[18:44] <wsirccc> well my package is qemu-buro. It is for that use and type of user, not much problem in designing that interface. One Person Buro.
[18:45] <wsirccc> sarnold
[18:45] <wsirccc> sarnold
[18:45] <wsirccc> sarnold
[18:45] <wsirccc> user is set to create is own image templating
[18:45] <wsirccc> pardon
[18:45] <wsirccc> sarnold:?
[18:45] <wsirccc> is/its
[18:46] <teward> give him a minute to read lol
[18:46] <wsirccc> lol about nothing?
[18:46] <teward> he's busy too you know :P
[18:46] <teward> we all are :P
[18:48] <wsirccc> teward: Thats no reason to laugh loud. Especially if you agree on the sake.
[18:49] <wsirccc> So one can hold fast: I have been understand falsely: My tool should imply:user is set to create is own image templating
[19:04] <wsirccc> could distribute raw images and virsh command line scripts to import the VM but assuming people have a working libvirt install may be asking a lot too
[19:04] <wsirccc>  could distribute raw images and virsh command line scripts to import the VM but assuming people have a working libvirt install may be asking a lot too
[19:04] <wsirccc> sarnold
[19:04] <wsirccc> you certainly could distribute raw images and virsh command line scripts to import the VM but assuming people have a working libvirt install may be asking a lot too
[19:05] <wsirccc> " a working libvirt install "
[19:05] <wsirccc> pardon
[19:05] <wsirccc> what is that?
[19:05] <wsirccc> sarnold:?
[19:13] <lordievader> !patience | wsirccc
[19:14] <wsirccc> ok
[19:24] <sarnold> wsirccc: take a look at the virsh manpage for the full range of options available
[19:26] <wsirccc> sarnold: did that some time ago, what should I learn from it?
[19:26] <Blueking> just wonder about dhcp lease time  dhcp request, expire , offer, nak ack   My ethernet interface has mac address  00.25.90.aa.b3.b4  BUT on dhcp logs it says its mac are 00:25:90.aa.bf.97    why is it difference from my ethernet interface ?
[19:29] <Blueking> sarnold u've seen something like that ?
[19:29] <sarnold> Blueking: no, sorry, doesn't make any sense to me
[19:30] <sarnold> wsirccc: virsh shows how to import vms from files into libvirt
[19:30] <Blueking> trying to figure out why renew of ip lease fails
[19:42] <Blueking> patdk-lap alive ?
[19:43]  * Blueking pinching on patdk-lap's right eyeball,,, could it be fake ?
[19:46] <patdk-wk> heh?
[19:46] <patdk-wk> crappy dhcp server
[19:46] <Blueking> think it's dhclient giving wrong mac address somehow
[19:46] <patdk-wk> how is that possible?
[19:47] <patdk-wk> dhclient doesn't give a mac
[19:47] <Blueking> it doesn't ?
[19:47] <patdk-wk> nope
[19:47] <Blueking> tcpdump + dhcp.pcap file shows both wrong mac address
[19:47] <patdk-wk> you do know your osi stack right?
[19:48] <Blueking> when dhcp server was working as expected 3 times in row it used RIGHT mac address 2 days ago
[19:48] <patdk-wk> ya, not helpful
[19:48] <patdk-wk> as I cannot see it
[19:48] <Blueking> I have it in logs
[19:49] <Blueking> trying to understand where this wrong mac coming from :/
[19:49] <Blueking> can I pm u ?
[19:49] <patdk-wk> no
[19:49] <Blueking> just show you the two mac addresses  nothing else
[19:50] <jrwren> a mac comes from a NIC
[19:50] <jrwren> dhcp server assigns IP to a MAC (poorly worded)
[19:51] <patdk-wk> it cannot cross l2 segments
[19:51] <Blueking> right mac  00:25:90:aa:b3:b4    wrong mac   00:25:90:aa:bf:97
[19:51] <nacc> Blueking: so you have a DHCP server, and you have your client, and your client has MAC 00.25.90.aa.b3.b4, but your server says it gave a lease to 90.aa.bf.97 ?
[19:51] <nacc> and no mention of b3.b4?
[19:51] <Blueking> mine ISP's dhcp server
[19:51] <Blueking> yes
[19:51] <patdk-wk> isp's do not do l2 segments
[19:51] <patdk-wk> they use dhcp relays
[19:51] <patdk-wk> the dhcp relay is likely having issues
[19:52] <sarnold> same mac vendor is funny thuogh
[19:52] <patdk-wk> or maybe the isp supplied hardware, dsl/cable/... modem
[19:52] <sarnold> ahhh
[19:52] <sarnold> then they're all going to be arris or whatever :)
[19:53]  * patdk-wk is just happy grub got fixed
[19:53] <patdk-wk> and I can boot encrypted with uefi
[19:53] <sarnold> \o/
[19:53] <Blueking> toshiba  dhcp server ?
[19:53] <patdk-wk> Blueking, you aren't doing something odd are you, like using bonding/teaming/....
[19:54] <patdk-wk> or have some kind of shared nic (ipmi/ilo/....)
[19:54] <Blueking> dhcp server  on ip 81.167.184.1
[19:54] <Blueking> no, the other nic are busy on lan (lokal)
[19:54] <Blueking> local
[19:54] <Blueking> that has ip 10.25.0.1
[19:55] <patdk-wk> nothing you can locate on your network has that other mac?
[19:55] <patdk-wk> or one close to it
[19:55] <Blueking> just wonder why bf:97 instead of b3:b4
[19:56] <Blueking> the other nic b3:b5
[19:56] <nacc> Blueking: if this was your ISP providing a DHCP lease, what logs were you referring to above? do you mean DHCPREQUEST or so in /var/log/syslog?
[19:56] <Blueking> tcpdump.cap + tshark  udp port range 67-68
[19:57] <Blueking> both shows same wrong mac
[19:57] <nacc> Blueking: and i guess that goes back to patdk-wk's question -- is it possible another device on your network has that mac?
[19:57] <patdk-wk> what does both mean?
[19:57] <patdk-wk> what where in those does it show it?
[19:57] <Blueking> 2 days ago  there were 3  correct lease renew and IT used the correct mac address
[19:58] <Blueking> I only have one supermicro device   this mobo
[19:58]  * patdk-wk gives up
[19:58] <Blueking> wait I'll show pcap file
[19:59] <jrwren> i wonder if it is something silly like you rebooted and your ethernet devices swapped for some reason.
[20:02] <Blueking> em1 = 00.25.90.d4.c1.c6 and em2 00.25.90.d4.c1.c7
[20:02] <Blueking> em1 & em2 thernet devices
[20:03] <Blueking> dhcp server  think it's 00.25.90.d4.bf.97
[20:04] <patdk-wk> doubtful
[20:04] <Blueking> mine logs show it
[20:04] <patdk-wk> the dhcp server is not on the same l2 segment as you
[20:05] <patdk-wk> so there is no way you can see it
[20:05] <patdk-wk> you can see the dhcprelay/gateway that responds to you
[20:05] <wsirccc> sarnold: so "libvirt install" what does it for me, what I should have done otherwise manually.
[20:06] <Blueking> Internet Protocol Version 4, Src: 81.167.184.1, Dst: 255.255.255.255
[20:06] <wsirccc> sarnold:having read: http://wiki.libvirt.org/page/FAQ#What_is_some_of_the_major_functionality_provided_by_libvirt.3F
[20:07] <Blueking> Ethernet II, Src: SuperMic_d4:bf:97 (00:25:90:d4:bf:97), Dst: Broadcast (ff:ff:ff:ff:ff:ff)
[20:07] <Blueking> wireshark  on pcap file
[20:07] <Blueking> the correct one
[20:08] <Blueking> Ethernet II, Src: SuperMic_d4:c1:c6 (00:25:90:d4:c1:c6), Dst: ToshibaT_01:00:01 (00:06:00:01:00:01)
[20:08] <jrwren> Blueking: pastebin `ip addr show` output and `ip route show` output
[20:09] <jrwren> Blueking: and pastebin `ip neigh show` output for good measure
[20:09] <Blueking> http://paste.ubuntu.com/15678338/
[20:09] <Blueking> http://paste.ubuntu.com/15678342/
[20:10] <Blueking> http://paste.ubuntu.com/15678354/
[20:10] <patdk-wk> in no way whatsoever is that mac address from the dhcp server
[20:11] <Blueking> I have had problem with dhcp renew of ip/lease
[20:11] <Blueking> for 3 weeks
[20:12] <supNow> just upgraded ubuntu server to 14.04 and I'm getting a grub_term_highlight_color not found error. google search suggests it can't find grub but I havn't found any real information on how to fix. Does anyone know a quick fix to this issue?
[20:12] <Blueking> jrwren what u think ?
[20:13] <jrwren> Blueking: I think it is very strange.
[20:13] <patdk-wk> we still dunno if that mac is somewhere else on your network
[20:14] <Blueking> havn't noticed the mac address before 30 minute ago
[20:14] <patdk-wk> or even in another home
[20:15] <Blueking> dhcp.pcap showed wrong mac when dhcp renew lease fails... when dhcp server  finally 3 times in a row two days ago in 4-5 hours worked as it should do it used the right mac address
[20:16] <Blueking> I can upload  two pcap files that u can study in wireshark
[20:17] <wsirccc> sarnold:https://linuxcontainers.org/lxd/try-it/ this is amazing. But I still do not feel very secure that I am on the right way..
[20:17] <wsirccc> ..
[20:17] <jrwren> can you kill your running dhclient and run `dhclient -1 -v -pf /run/dhclient.em1.pid -lf /var/lib/dhcp/dhclient.em1.leases em1` ?
[20:17] <Blueking> ok
[20:17] <jrwren> Blueking: err, add a -d to that.
[21:17] <DirtyCajun> if i had a managed switch and i vlan separate vlan 1 to port1 (my server)  and port 2 (the modem).. then vlan2 port 1 (my server) and ports 3-8 .. i could dhcp with only 1 nic that way right?
[21:24] <Blueking> ok problem solved :)
[21:33] <Blueking> ok problem solved jrwren and patdk-lap patdk-wk :)
[21:46] <patdk-wk> till next week?
[21:46] <patdk-wk> what do you believe solved it?
[21:56] <NwS> heya guys, any suggestions for a full clone/backup of an ubuntu server 14.04LTS without a restart?
[21:56] <TJ-> NwS: image the block devices, or clone the installed packages and configs?
[21:59] <NwS> I want to make a full backup copy just in case something happens and restore everything later
[21:59] <NwS> And yeah have never done this before so -.-"
[22:01] <TJ-> !backups | NwS
[22:01] <NwS> ty for the info TJ-
[22:01] <Blueking> patdk-wk: ipmi   management lan  had same IP as ethernet inerface em1    ipmi have mac address 00.25.90.d4.bf.97
[22:01] <Blueking> TJ-  fixed problem
[22:02] <Blueking> TJ-  it was ipmi/management lan that was reason for net drop outs... dhcp did see ipmi's mac address
[22:02] <TJ-> Blueking: so I hear! So the problem was due to the server-board
[22:02] <Blueking> yes
[22:03] <Blueking> it's good to have it fixed :)
[22:03] <Blueking> TJ-  patdk-wk didn't believe in wrong mac reason :P
[22:04]  * Blueking pats patdk-wk on shoulder
[22:05] <TJ-> Blueking: so the IPMI was getting between the ISP and the OS?
[22:05] <Blueking> supermicro   for some years  had ipmi and ethernet nic on same physical connector ?
[22:05] <Blueking> but  not anymore says RoyK
[22:07] <Blueking> bedtime  hopefully there will be none disconnects for long time
[22:07] <Blueking> we'll see in some days :)
[22:11] <sarnold> Blueking: if you've only got one ethernet wire to your supermicro, it'll run the ipmi/bmc _and_ regular NIC functionality on that one NIC
[22:12] <Blueking> ok
[22:12] <Blueking> think I have to swap nic's  so ipmi are only visible on local network
[22:13] <Blueking> just tested arping -i em1  impi's IP
[22:13] <Blueking> ipmi
[22:13] <Blueking> not visible on em2
[22:17] <sarnold> nice, that's robably a good idea
[22:18] <sarnold> i'm not sure i'm going to bother at home, chances are good nsa / mossad are already embedded (cool dudes i'm sure)
[23:24] <rbasak> hallyn: I've filed bug 1567696 and bug 1567695