[00:19] <JanC> tikun: it's installed on both, I think, but boots from the ESP partition
[00:21] <tikun> hmm
[00:21] <tikun> kk, thx
[00:21] <JanC> I assume the other one is in case you switch to legacy boot
[00:24] <lucas_ai> I can't get my @reboot crontab job to run my program. There is no error message in syslog (it shows up as if the command was run but it's not running). Any ideas?
[00:28] <Kamilion> lucas_ai: are you expecting PATH to be sane in a crontab, or expecting a specific PWD?
[00:29] <lucas_ai> I'm doing this: @reboot /usr/local/bin/autorotate.py > /usr/local/bin/log.txt
[00:29] <lucas_ai> I don't think I need the path to be sane. Do I?
[00:30] <Kamilion> well, python certainly will be unhappy depending on what you try to import
[00:30] <lucas_ai> but I don't see any error message either
[00:31] <lucas_ai> How am I supposed to run a script whenever my ubuntu loads?
[00:31] <Kamilion> make a systemd unit?
[00:33] <lucas_ai> that means adding a lot of code. Why can't I just write my script command somewhere and that's it??
[00:34] <lucas_ai> upstart was fine but systemd is stupid
[00:35] <Kamilion> um
[00:35] <Kamilion> ...
[00:35] <Kamilion> I feel strongly opposite.
[00:35] <Kamilion> I haaaaaaaaaaaaaaaaaaaaaaaate upstart with a passion.
[00:36] <Kamilion> trying to write units in it and get them to behave properly and in order was nothing but trouble
[00:36] <lucas_ai> making a unit requires lots of writing. upstart if I remember correctly is just almost a regular bash script
[00:36] <Kamilion> no
[00:36] <Kamilion> upstart's format is almost the same as systemd.
[00:36] <lucas_ai> maybe I'm talking about an even older daemon system
[00:36] <Kamilion> https://github.com/kamilion/kamikazi-core/tree/master/resources/xenial/mods/etc/systemd/system
[00:36] <Kamilion> here are three service files.
[00:37] <lucas_ai> I just want to add my script to be run on start. why is it so hard?
[00:37] <Kamilion> it isn't anymore.
[00:37] <lucas_ai> I don't wanna make a unit script for every bash/python script I ever want to run on startup. This is disappointing
[00:37] <Kamilion> https://github.com/kamilion/kamikazi-core/blob/master/resources/xenial/mods/etc/systemd/system/disque.service
[00:37] <Kamilion> look dude, it's ten lines.
[00:38] <lucas_ai> the command in that script is only one line
[00:38] <lucas_ai> I don't wanna memorize the whole unit syntax
[00:38] <lucas_ai> where do I just write "/usr/local/autorun.py" and that leaves it running since it boots?
[00:39] <Kamilion> [Unit]
[00:39] <Kamilion> After=syslog.target
[00:39] <Kamilion>  
[00:39] <Kamilion> [Service]
[00:39] <Kamilion> Type=oneshot
[00:39] <Kamilion> ExecStart=/usr/local/bin/autorotate.py > /usr/local/bin/log.txt
[00:40] <Kamilion> or drop the whole [Unit] section if you don't need it ordered at boot.
[00:40] <Kamilion> generally a good idea to wait until at least syslog or network is running though
[00:40] <lucas_ai> awesome. Do I need the Type line?
[00:40] <Kamilion> yeah, unless you want it to default to 'restart'
[00:41] <lucas_ai> what does that mean?
[00:41] <Kamilion> systemd will try to restart it if it exits cleanly.
[00:41] <Kamilion> if it throws an error more than 3 times, it'll be put in the 'failed' state and not restarted.
[00:41] <Kamilion> oneshot just launches it once, like the old systemv init scripts did.
[00:41] <lucas_ai> lol
[00:41] <Kamilion> 'just run this, GTFOMW'
[00:41] <Kamilion> which is pretty much what you're asking for.
[00:42] <lunaphyte> that's all well and good, but cron @reboot should work, and there's nothing wrong with it.  it's perfectly fine
[00:42] <Kamilion> oneshot is what you need for it to run and become a 'SUCCESS'
[00:43] <Kamilion> eh, I've never gotten cron to work properly on @reboot jobs
[00:43] <lucas_ai> should I put this unit in /etc/init.d/ ?
[00:43] <Kamilion> the unit file?
[00:43] <lucas_ai> yes
[00:44] <Kamilion> /etc/systemd/system/mylogrotate.service
[00:44] <Kamilion> systemctl daemon-reload
[00:44] <Kamilion> systemctl start mylogrotate
[00:44] <Kamilion> systemctl status mylogrotate
[00:44] <Kamilion> check the log messages it prints
[00:46] <lucas_ai> sudo service autorotate start Job for autorotate.service failed because the control process exited with error code. See "systemctl status autorotate.service" and "journalctl -xe" for details.
[00:48] <lunaphyte> it's pretty easy to troubleshoot @reboot.  delete /run/crond.reboot, then restart cron and watch the logs
[00:48] <lucas_ai> I need to run my script as root. How could I do that in my unit file, Kamilion
[00:50] <Kamilion> you didn't ask for it to run as any other user, so it'll run as root.
[00:51] <lucas_ai> then why am I getting an error in my script? Sep 11 17:47:41 lucas-Surface-Pro-3 autorotate.py[10820]: Can't open display
[00:52] <Kamilion> that sounds like it's trying to access the xorg display somehow.
[00:53] <lucas_ai> yes. When i run it manually, it works just fine
[00:55] <Kamilion> if you're trying to start a GUI program on boot, you're better off using the xdg autostart then. http://askubuntu.com/questions/63407/where-are-startup-commands-stored
[00:55] <lucas_ai> wtf. This is a command line script. But apparently it needs to get a screen
[00:56] <Kamilion> eh, then cron would have never worked in the first place.
[00:57] <Kamilion> After=graphical.target
[00:58] <Kamilion> that will start the script after a display's available; but depends on further configuration such as lightdm autologin or something to actually have the desktop at a meaningful place to execute.
[01:01] <lucas_ai> lol
[01:01] <lucas_ai> I made a .desktop file in /etc/xdg/autostart
[01:01] <lucas_ai> how can I test if it works?
[01:01] <Kamilion> double click it?
[01:01] <Kamilion> i assume you just right clicked to make the desktop file
[01:01] <Kamilion> but I have no idea what kind of workflow you have or how much knowledge
[01:01] <lucas_ai> no. on the terminal
[01:02] <Kamilion> oh, I have no idea how to use .desktop files on the terminal
[01:02] <lucas_ai> i'll try in nautilus then
[01:02] <Kamilion> they're the equivolent to windows shortcut icons
[07:46] <ducasse> i need to reinstall a 16.04 zfs file server, do i just export the pools on the old install and import on the new, or do i also copy zpool.cache over?
[07:48] <sikun> I really need to learn zfs
[07:53] <beardface> ducasse: exporting and importing should do it
[07:54] <ducasse> beardface: ok, thank you :) i'm still learning zfs, and don't want to mess up my pools :)
[09:32] <psjbeisler> anybody have experience with containers yet, specifically LXD not Docker
[09:32] <psjbeisler> or KVM
[09:44] <KlausedSource> does anyone know of a tool with that i can extract information (page dimensions, number of sheets/sides) from a pcl print file?
[10:00] <jonah> hi can anyone help with a weird problem. I have 3 servers. A can ssh to B and C. B can ssh into A and C. But C can only ssh into B not A. I can't figure out why!?
[10:02] <lordievader> jonah: Check the auth log.
[10:02] <lordievader> That makes a good start.
[10:03] <cpaelzer> also ssh can get much more verbose what is actually going on with -v -vv -vvv
[10:03] <cpaelzer> in case you find nothing on the auth log and need to look at the clients side what might be wrong
[11:26] <jamespage> Odd_Bloke, morning
[11:26] <Odd_Bloke> jamespage: o/
[11:27] <jamespage> Odd_Bloke, when will the next yakkety daily images be published?
[11:27] <jamespage> I'm trying to get some yakkety/openstack testing back up and running - I'm working around the apt-mirror issue (by setting it)
[11:27] <jamespage> but our current daily (from the 3rd) breaks on upgrade of grub
[11:28] <Odd_Bloke> jamespage: We're having some issues testing on ppc64el which are breaking things, which I need to dig in to; I'll manually override the latest one though, as it's been a while since we had one.
[11:28] <jamespage> | 4e84df6a-72fc-41f2-b839-41430dacd6e6 | ubuntu-daily/ubuntu-yakkety-daily-amd64-server-20160902-disk1.img             |
[11:28] <jamespage> Odd_Bloke, pls
[11:30] <Odd_Bloke> jamespage: That's in progress now, you should see it show up in the next few hours.
[11:30] <jamespage> Odd_Bloke, \o/
[11:31] <Odd_Bloke> And now I guess I should dig in to these ppc64el failures, as my "wait a week" strategy hasn't yielded the results I was hoping for. :p
[11:36] <Ussat> o/
[12:04] <cpaelzer> jamespage: fyi I'm providing an ppc64el dpdk enabled clone of openvswitch in a ppa
[12:04] <cpaelzer> jamespage: along a ppc64el enabled dpdk of course
[12:06] <cpaelzer> jamespage: once 17.04 opens up I might get to you to incooperate the changes - but that to some extend depends on the further work of IBMers related to that
[12:26] <Ussat> cpaelzer, any IBM'er in particular I can prod, I used to work there
[13:23] <neoark> http://legalhackers.com/advisories/MySQL-Exploit-Remote-Root-Code-Execution-Privesc-CVE-2016-6662.html
[13:28] <Pici> neoark: #ubuntu-security would be a good place to share that
[13:30] <mdeslaur> neoark: we'll be publishing updates for that soon
[13:34] <mdeslaur> rbasak: we need to update mysql to 5.7.15 ^
[13:36] <rbasak> ack
[13:56] <jamespage> ddellav, if you want todo some xenial/newton testing with https://launchpad.net/~james-page/+archive/ubuntu/newton/+packages overlaid post deploy that would be helpful
[13:56] <jamespage> cpaelzer, ok - I know the 6wind team did some work for IBM on power
[13:57] <ddellav> jamespage i'd love to help. Do you mean deploying newton on xenial and using that ppa as a target in the next.yaml?
[13:59] <jamespage> ddellav, no - you need to deploy either the xenial-newton or xenial-newton-branch targets from o-c-t
[14:00] <jamespage> and then add and upgrade that ppa - I'd use juju run  for that
[14:00] <jamespage> its basically a rev to the virt stack - want to de-risk it a bit
[14:08] <jamespage> smb: hey do we need libvirt and libvirt-python to minor version match? we currently have 2.1.0 and 2.0.0 in archive
[14:09] <smb> jamespage, I cannot say for 100% but Debian had the same minor version delta last time I looked
[14:10] <jamespage> smb, hmm
[14:11] <jamespage> we've had issues before with unpaired versions
[14:11] <smb> jamespage, now its even 2.2/2.1 against 2.0
[14:11] <jamespage> i.e. new features not being exposed
[14:50] <mdeslaur> rbasak: fyi, for xenial 5.7.15, I'm going to use the new upstream tarball that contains the embedded boost
[14:51] <rbasak> mdeslaur: I believe we're already doing that.
[14:51] <rbasak> mdeslaur: (so that should be fine)
[14:52] <mdeslaur> rbasak: oh! cool, I didn't notice that. thanks
[14:54] <rbasak> mdeslaur: the watch file should still work. I hope :)
[14:54] <mdeslaur> heh, I should have tried it :)
[16:05] <hariom> I have opvn file from my old openvpn configuration. I want to setup vpn connection again. Where should I keep this file? I have compiled and installed open vpn 2.3.4
[18:04] <EmilienM> jamespage, coreycb: hey oh
[18:04] <EmilienM> I still see some failures when deploying neutron/linuxbridge https://review.openstack.org/#/c/368826/
[18:04] <EmilienM> (scenario003 is using linuxbridge backend) - not sure you updated everything yet
[19:13] <jamespage> EmilienM, all of b3 is there for core projects - http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/newton_versions.html
[19:13] <jamespage> including the bits for loading modules and stuff
[19:14] <EmilienM> jamespage: mhh, ok
[19:22] <jamespage> EmilienM, oddly enough I noted a few tempest failures pm today when doing some ovs testing
[19:22] <jamespage> might be related
[19:22] <jamespage> related to instances not being accessible
[19:22] <EmilienM> with ovs we have 0 issue
[19:22] <EmilienM> only with linuxbridge
[19:24] <jamespage> EmilienM, that test (test_network_basic_ops.TestNetworkBasicOps) is the only one in the tempest test suite that actually verified floating-ip -> instance connectivity I think
[19:24] <EmilienM> yep, that is the one that failed for us when using linuxbridge backend in neutron
[19:24] <EmilienM> when using ovs, it pass fine
[19:26] <jamespage> EmilienM, you can see that the instances does not get an IP address via dhpc
[19:27] <jamespage>  Sending discover...
[19:27] <jamespage> multiple times
[19:29] <jamespage> ddellav, coreycb: https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1622715
[19:29] <jamespage> I've tripped over that fairly reliably
[19:30] <ddellav> jamespage I'm having an issue starting up denial-newton on serverstack due to neutron, i'm still debugging but it might be related to that bug as well
[19:30] <ddellav> *xenial-newton
[19:30] <jamespage> ddellav, symptomatically you can't access instances
[19:30] <jamespage> ddellav, the status of the router port is DOWN
[19:30] <jamespage> until the n-ovs-agent on the n-gateway unit is restarted
[19:31] <ddellav> ok, i'll try that. So far I'm just seeing an error on config-changed during startup. I'm digging through logs but i don't see much. I'm redeploying now
[19:33] <jamespage> EmilienM, are the puppet modules using the native ryu support for managing ovs, or the older ovs-vsctl driver?
[19:35] <EmilienM> jamespage: we use the one in neutron upstream default
[19:35] <jamespage> EmilienM, ok so ryu for newton
[19:36] <EmilienM> yep
[20:13] <jamespage> ddellav, doing a debug enabled run to see if I can figure out that n-ovs problem
[20:14] <jamespage> ddellav, -o debug=True with deployer btw
[20:46] <sarnold> cpaelzer: congratulations :)
[20:48] <cpaelzer> thank you sarnold
[21:30] <jamespage> ddellav, Ithink that neutron problem is caused by ryu
[21:30] <jamespage> bumped to 4.5 which has some eventloop exception fixes
[21:47] <Kevin> Is there a way I can autoconfigure my network without reinstalling the OS?
[21:50] <jamespage> ddellav, coreycb: ok so after some testing
[21:50] <jamespage> ovs 2.6 + libvirt 2.1.0 appears functional afaict
[21:50] <jamespage> with newton xenial
[21:51] <jamespage> libvirt 2.1.0 will require some charm updates for nova-compute for the libvirt-bin -> libvirtd daemon rename
[21:52] <DzAirmaX> hi all
[21:52] <DzAirmaX> can someone help me to set up kdump properly? I have installed it  and everything seems fine but during the last crash I couldn't find any logs in /var/crash
[21:55] <compdoc> MySQL Hit By "Critical" Remote Code Execution 0-Day
[21:58] <coreycb> jamespage, that's good
[22:17] <DzAirmaX> hey
[22:17] <DzAirmaX> someone has already experienced crash with no kdump log?
[22:41] <RoyK> DzAirmaX: that happens, usually with something funny in the hardware. Start with a memory test
[23:36] <keithzg> Any suggestions for fully-compatible PCIe SATA controllers? (Preferrably on the more affordable side, but any suggestions would be welcome.)
[23:36] <keithzg> The one I've got in a new server seems . . . questionable, although it's hard to pin down quite what's going wrong.
[23:38] <compdoc> I use LSI controllers, but with the bios changed to get rid of the raid features
[23:39] <compdoc> not very expensive on ebay, and support sata and sas
[23:56] <keithzg> compdoc: Thanks for the suggestion. By "with the bios changed" do you mean just flipping something in their firmware settings, or outright flashing modified firmware on them, or?
[23:59] <DzAirmaX> RoyK: how long shoud I run memtest for being sure everything is OK?