[02:05] <smoser> are the values for dhcp really 'no' and 'true' ?
[02:05] <smoser> https://netplan.io/examples
[02:06] <smoser> i guess i tuyped too quickly. the values there in the examples are 'no', 'yes', and 'true'.
[02:07] <smoser> probably should make that consistent.
[02:08] <mason> smoser: There's a #netplan channel, FWIW. Might be worth talking about it there too.
[02:09] <sarnold> is that a general fault of yaml? yes/no on/off true/false?
[02:15] <coreycb> jamespage: we're mostly caught up with stein packages
[02:16] <sarnold> smoser: yes/no on/off true/false may just be a general fault of yaml
[05:14] <jamespage> coreycb: awesome
[06:22] <jamespage> coreycb: I'm intending on working through snapshots for any that lack b's
[06:22] <jamespage> nova revealed a new dep/refactoring thing - os-resource-classes
[06:22] <jamespage> uploaded for aa review
[07:43] <andre144k> hi all, how to kill a linux-prozess which willnot stop via "killall -9 <name>" and willnot stop via "kill -9 <pid>"
[08:04] <jamespage> mwhahaha: I've uploaded a keystone snapshot for stein - should be into proposed in the next few hours
[08:39] <lordievader> Good morning
[08:54] <frickler> jamespage: coreycb: we found this issue when upgrading the OpenStack CI to bionic, looks like a packaging issue to me https://bugs.launchpad.net/cinder/+bug/1819819
[14:01] <jamespage> sahid: looking at your openstack pkg updates now
[14:33] <sahid> jamespage: thanks, let me know or comment on the doc if you need me i change something
[14:34] <jamespage> sahid: all looks ok so far - glance uploaded to ubuntu
[14:37] <sahid> ack thanks
[14:43] <jamespage> frickler: commented - there is a mix of package based and pip installed rtslib-fb on the log output
[14:44]  * blackflow looks at bug #1797386 and sends some positive vibes toward everyone involved ...    <3
[14:45] <frickler> jamespage: oh, good point, I didn't notice that, thx
[14:47] <frickler> jamespage: I tested locally but only after devstack had run, so probably had the same pip installation state
[14:47] <jamespage> frickler: most likely
[14:48] <jamespage> frickler: my experienced ubuntu bug triage eyes automatically zone in on anything starting with /usr/local ;)
[14:52] <jamespage> sahid: build failure on https://launchpadlibrarian.net/415128633/buildlog_ubuntu-disco-amd64.mistral_8.0.0~b2-0ubuntu1~ubuntu19.04.1~ppa201903141431_BUILDING.txt.gz
[15:12] <sahid> jamespage: hum... weird they pass locally with pbuilder
[16:13] <jamespage> sahid: looks like a racey test - don't worry about it
[16:58] <jamespage> sahid: I think placement needs a dependency on python-os-resource-classes
[17:05] <jamespage> sahid: ok all done apart from placement - python3-os-resource-classes was accepted into disco this morning
[17:05] <jamespage> so you should be good to go on that one.
[18:19] <geigerCounter> I have a question: Is there any reason to change my hostname through the cloud.cfg?
[18:22] <geigerCounter> My VPS provider had it configured this way, but when I updated I switched to the new config and manually edited in their change. I'm just curious about it.
[18:25] <blackflow> geigerCounter: change from what to what? one reason would be if you're running a mail server and want to ensure proper rDNS
[18:25] <blackflow> (that rDNS matches the forward resolution)
[18:26] <geigerCounter> blackflow: To the address of the VPS in their system, I guess. It's vpsXXXXXX.vps.ovh.ca, since I'm using OVH. ( With the X's being my VPS number. )
[18:28] <blackflow> geigerCounter: isn't that the default? Aren't their VPS images configured that way?
[18:30] <geigerCounter> Yes, it is the default for their VPS images. I was doing a manual update and it asked me if I wanted to install the package maintainer's version. After looking through the differences I did so and edited the config OVH did back in. I'm just here asking about why it is this way for my own education.
[18:31] <blackflow> well I understood you were asking if there's a reason to change from default to something else, and I gave you one. It really boils down to how do you want your server to be named.
[18:32] <geigerCounter> I see. I suppose I'd have to ask them if it's even allowed to change it. I would like to change it to my public DNS address.
[18:32] <geigerCounter> I am running a mail system, for one thing.
[18:33] <geigerCounter> Could that be a reason why Gmail's not sure about my mail?
[18:34] <geigerCounter> I'd used a mail configuration testing service that reported no problems, but Gmail still issues a warning with mail sent from my server.
[18:38] <blackflow> geigerCounter: you don't have to ask OVH, you can chage it to whatever you want. Note that there's to facets of that. One is the hostname configured onthe machine itself. The other is rDNS for its IPs, configurable in OVH's panels.
[18:39] <blackflow> you can only affect your local hostname on the machine itself. it's the name various systems get when they ask the OS "hey, what's my hostname here".
[18:40] <blackflow> GMail will complain if you don't have rDNS, and it's wise to have rDNS and hostname the same.
[18:41] <geigerCounter> I see. I'll set it to what I need and then I'll check out the rDNS in my OVH panel. While I'm here, does anyone have a link to a good guide for setting up a startup script to run a program as a service on startup? ( How to write an init.d script, I guess? )
[18:42] <blackflow> geigerCounter: it's a systemd unit file these days, far more simple than an init shell script. unless you're still running 14.04 or 12.04?
[18:43] <geigerCounter> I think I'm running 18.04, but I'm at least running 16.04. So yeah.
[18:44] <geigerCounter> I just haven't really done it before. I need to do so though, because I have too many things running on my VPS that I have to start manually on a reboot.
[18:44] <blackflow> geigerCounter: ton of articles on on google. for example start here   https://www.digitalocean.com/community/tutorials/understanding-systemd-units-and-unit-files
[18:44] <geigerCounter> Thanks.
[18:46] <geigerCounter> Is it possible to start one of these in interactive mode with screen or something so I can just su to the acc that's responsible for running it and interact with the console? ( I ask, because I have a Minecraft server. Worse case scenario, I can remote admin it through IRC or Discord with a plugin I have if it doesn't fail to launch when not in interactive mode. )
[18:50] <blackflow> geigerCounter: sure   https://superuser.com/questions/1276775/systemd-service-python-script-inside-screen-on-boot
[18:51] <blackflow> you can start anything from systemd, it's just the question of what, and how (service "Type").
[18:51] <geigerCounter> I see! Very nice. Thank you so much.
[18:54] <geigerCounter> Alright, so back to asking about reverse DNS since I've never really done anything with DNS other than registering my domain name and setting up the appropriate pointers to my IP and an SPF record - what do I want to set my reverse DNS to in my OVH panel? My domain name or something else?
[23:47] <geigerCounter> I haven't been able to find the info I'm looking for - does Systemd always kill child processes when you stop a service?
[23:50] <tomreyn> i don't actually know, but i would hope so. what makes you ask?
[23:52] <tomreyn> geigerCounter: ^
[23:55] <geigerCounter> tomreyn: Well I mean, yeah I guess it's the usual and sane behaviour, but I have a weird use case. I'm trying to make a Systemd Unit that launches a Minecraft Spigot server in a screen with the appropriate user I set up for that. That way all I have to do is su to that user and reattach to manage the console and the server will start on boot and can be controlled as any other system process. The thing
[23:55] <geigerCounter> is getting it to gracefully shut down. Right now it's pretty brutally killed and that's most likely bad for the health and longevity of the server.
[23:56] <geigerCounter> So the chain of events from systemd goes: Systemd -> screen -> BASH script -> Java running the spigot jar.
[23:57] <geigerCounter> What I want to have happen on calling "service spigot stop" is that it uses screen to stuff a 'stop' command and wait for the java process to end normally as the server shuts down and then exit up from there with screen terminating as expected.
[23:57] <geigerCounter> It kills the java process instead.
[23:59] <geigerCounter> The only thing I can think to do to get around this is to make a bash script which stuffs the 'stop' to the java console and waits five minutes before exiting and set TimeoutStop to five minutes also.
[23:59] <geigerCounter> But is there a better way?
[23:59] <geigerCounter> ( Systemd novice here. )