18:03:34 <julen> #startmeeting training-labs
18:03:35 <openstack> Meeting started Mon Aug 22 18:03:34 2016 UTC and is due to finish in 60 minutes.  The chair is julen. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:03:36 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:03:39 <openstack> The meeting name has been set to 'training_labs'
18:04:11 <julen> #topic rst parser
18:04:25 <julen> rluethi: did you give it a try?
18:04:38 <rluethi> I was about to ask you the same :)
18:04:42 <rluethi> yes, I did.
18:05:01 <julen> hehe... I was talking with Pranav on Friday, and he showed me how it works
18:05:17 <julen> by the way... were is our artist?
18:05:24 <rluethi> It's not yet quite as user-friendly as I would like it to be.
18:05:29 <julen> #ping dbite
18:05:39 <rluethi> he sent an email, he's still at work.
18:05:48 <rluethi> I just barely made it :)
18:06:07 <julen> oh... I didn't get it
18:06:31 <rluethi> he must have been in a hurry, he just sent a quick note.
18:06:39 <julen> well... the cake is pretty done, now it just needs the cherry on top, so the formatting
18:07:10 <julen> well.. first things first
18:08:17 <rluethi> I would like to think that it will be ready for Newton, but I'm afraid it won't be there in time.
18:08:23 <julen> I really like the way he handled the different types of blocks, did you take a look to the code?
18:08:57 <julen> I think there's not much left... just to output the stuff in bash format, instead of lists
18:09:01 <rluethi> I looked at it to check why the code needed to import attr.
18:09:16 <julen> oh! also... what do you think about adding some command line arguments and checks?
18:09:59 <rluethi> Pranav has definitely planned something like this, he just didn't get around to adding it.
18:10:03 <julen> I still don't understand that attr thing
18:10:19 <julen> we were talking about that
18:10:24 <rluethi> Yes, of course it needs at least a "--help" or something.
18:11:14 <julen> in principle, it shouldn't need it, in the context of osbash, but it's always good to have at least a "-h" and check that there is a valid input file
18:11:46 <rluethi> of course it needs it, I had no idea how to use it when I wanted to test it.
18:11:57 <julen> also for debugging... imagine that we accidentally pass an empty variable or something
18:11:59 <rluethi> that's all the use case I need.
18:12:07 <julen> hehe! me neither :)
18:13:05 <rluethi> so, I tried to help, but quite frankly I don't know what to test and how, or what to look for.
18:13:31 <julen> I think that the ball is in his roof right now
18:14:12 <julen> we could try to format those lists as bash, but he probably has a good draft almost ready to push, so it could be a waste of time
18:14:13 <rluethi> I do know that pylint almost fainted, but that's mostly over variable names and I don't care about that all that much.
18:15:08 <rluethi> We should probably start looking at Newton. It will be a manual port.
18:15:20 <rluethi> It's usually several days worth of work.
18:15:29 <rluethi> Two days if all goes well.
18:15:32 <rluethi> It never does.
18:16:21 <julen> uff... these things always seem more simple in my head, but in the reality there are always more difficulties
18:16:35 <julen> I guess we'll have to wait until next week
18:16:55 <rluethi> okay.
18:16:59 <julen> by the way, I won't be here next week
18:17:44 <rluethi> so for the time being, you will work on rst2bash?!
18:18:41 <julen> well.. that's Pranav's thing. I can surely test and give feedback, or maybe add the command line arguments, but that's quite simple
18:19:07 <julen> he told me that I could take a look to the Newton update
18:19:18 <rluethi> anything else you are working on?
18:19:26 <julen> but I also have the backport you suggested in the queue...
18:20:09 <julen> mostly working on trying to get some time to take care of that backport you suggested last week :P
18:21:04 <julen> I guess I can find a few hours this week to take care of that
18:21:13 <rluethi> cool
18:21:13 <julen> what about you?
18:21:26 <rluethi> I am still working on the Python port.
18:21:31 <julen> (at least the backport.. the Newton thing will take longer)
18:21:43 <rluethi> And I ran into a problem with KVM and PXE.
18:21:44 <julen> uff.. that's another big one
18:21:57 <julen> #topic python-port
18:22:42 <rluethi> KVM needs to boot from the second NIC, but after trying the first NIC (unsuccessfully), it somehow keeps the first network as the one containing next-server.
18:22:50 <julen> one thing for both of you: could you put copy/paste kind of commands for testing those scripts?
18:23:02 <julen> (on the README)
18:23:11 <rluethi> which scripts?
18:23:25 <julen> both the python port and the rst script
18:23:44 <rluethi> python port: ./st.py --help
18:23:53 <julen> I tried to give a try to the python script and I think it didn't work for some reason (maybe dependency missing or something)
18:24:15 <julen> aha! ok :)
18:24:31 <rluethi> if it has any dependencies, that's most likely a bug.
18:24:49 <rluethi> it should not have any dependencies other than python, that is a design requirement.
18:25:04 <julen> I'll try it again and send you were did I get stuck exactly
18:25:47 <rluethi> I'm afraid the code in the repo is somewhat out of date, I'll have to push an update. I'll let you guys know.
18:26:08 <julen> that NIC issue... what do you mean with booting from the NIC?
18:26:32 <julen> for installing the operative system?
18:26:43 <rluethi> well, with VirtualBox I can specify the boot order of network interfaces.
18:26:54 <julen> oh! true...
18:27:00 <rluethi> so it starts PXE booting on the management network.
18:27:12 <rluethi> on KVM, I don't know how to change the order.
18:27:44 <rluethi> so it tries the wrong one first and it looks like it gets confused by that even if no server replies to those requests.
18:27:56 <julen> uff.. I don't remember it either...
18:28:49 <julen> you run it with virt-install right?
18:28:58 <rluethi> yes.
18:29:04 <rluethi> but that's not a requirement.
18:30:03 <julen> aha! here it is:
18:30:10 <julen> #link http://stackoverflow.com/a/19162267/5820360
18:30:58 <julen> oh.. but it does not specify which NIC...
18:31:26 <rluethi> yes, it talks about the order of different media, but not the order of NICs.
18:32:40 <julen> the problem is, that it probably requires hacking the xml file. That's not so elegant
18:33:00 <rluethi> bah. nothing in libvirt is ever elegant.
18:33:03 <rluethi> it's by design.
18:33:54 <julen> it is possible to select it with the GUI (virt-manager) so there should definitely be a way to do it
18:33:58 <rluethi> We already build XML files on the fly to configure KVM networks.
18:34:33 <julen> true... I did it the same way for bomsi
18:34:46 <rluethi> Interesting. I guess all we have to do is check the differences the switch causes in the XML files and do that.
18:36:11 <julen> that's a good idea! ... I was thinking of how frightening must be to start checking the source code of virt-manager... but that's much more simple :)
18:36:45 <rluethi> I sure hope so.
18:38:02 <rluethi> can you send me a screenshot of the option in virt-manager? I can't find it, but I'm afraid it may be because the Linux distro on my workstation is showing its age.
18:38:54 <julen> probably it is possible to pass some other tag, for example <boot dev="network" NIC:"MAC">
18:40:30 <rluethi> maybe.
18:40:58 <julen> ok, but first try: Boot options > Boot device order , and that should show Virtio Disk, IDE CDROM, NIC: MAC and NIC:MAC
18:41:04 <julen> don't you have that?
18:41:46 <rluethi> I don't think I do.
18:42:00 <rluethi> I have version 0.9.1
18:42:24 <julen> where should I put the screenshot?
18:42:37 <rluethi> just shoot me a mail.
18:42:39 <julen> isn't there some pastebin like thing for pictures?
18:43:25 <rluethi> yes, of course there's email, but don't you have something more complicated? :)
18:43:34 <julen> done
18:43:56 <julen> :P
18:44:05 <rluethi> awesome. thank you.
18:44:23 <rluethi> do we want to talk about the summit real quick as well?
18:45:01 <julen> I'm using virt-manager v 1.0.1
18:45:14 <julen> #topic summit
18:45:15 <julen> sure
18:45:49 <julen> let's cross our fingers for Pranav getting his talks, including the training-labs one...
18:46:16 <rluethi> So, Barcelona is a city constantly overrun with tourists. If we want to get hotel rooms in a decent location for a decent price, we should consider booking soon.
18:46:34 <julen> I totally agree
18:47:12 <julen> the thing is, I am starting next week on my new job, and in principle I think the would allow me to go, but I'm not 100% sure
18:47:33 <rluethi> the time to ask them is now :)
18:48:04 <rluethi> tell them the training-labs team is counting on you.
18:48:40 <julen> let's say, I am kind of 85% sure that it will work, but there are some uncertain variables, like the strong bureaucracy...
18:49:07 <rluethi> got a job at IBM? :)
18:49:14 <julen> sure.. I will know at latest next week
18:49:20 <julen> hehe!
18:49:47 <julen> well... Fraunhofer :P
18:49:56 <rluethi> Heh. Close enough.
18:50:04 <julen> hehe...
18:50:12 <rluethi> Although some of these science guys are pretty free wheeling.
18:51:13 <rluethi> Also, whether or not we get a talk, we have to arrange meetings with the docs people so we can sync up with them.
18:51:31 <julen> sure... this is why I am so excited about that job... but it's also true that there is usually plenty of paperwork, and well... think about this new guy who wants to go on a business trip (or vacation) right after starting...
18:51:37 <rluethi> We need to make sure that we understand their needs and that they understand how we can help them.
18:51:43 <julen> definitely
18:51:58 <julen> and it would be good to have some presence in the academy
18:52:36 <rluethi> you can do training courses at Fraunhofer :)
18:52:42 <julen> I still don't understand why they are suggesting newbies to get started with the developer tools...
18:52:56 <rluethi> who does?
18:52:57 <julen> hehe.. I can try...
18:53:55 <julen> well.. check places like ask.openstack, every time someone asks how to get started, there always always people suggesting the devsolution and the packsolution...
18:54:26 <rluethi> in other words: training-labs marketing sucks.
18:54:40 <rluethi> we need to get better.
18:54:45 <julen> so, I think that the openstack academy is certainly a place to go, and try to mention training-labs as much as possible
18:54:52 <rluethi> where's our mascot?
18:55:02 <julen> well... but it does work pretty good already
18:55:08 <rluethi> that's a good idea.
18:55:20 <julen> wait.. before the maskot...
18:55:34 <julen> hmm... what did I want to say...
18:56:18 <julen> oh! vbrownbag, if it doesn't work with Pranav's proposal, we should definitely go for a vBrownBag talk
18:56:58 <julen> oh! almost 21:00
18:57:17 <julen> let's talk about the maskot thing some other time
18:57:21 <rluethi> brown bag might be an option.
18:57:28 <julen> but I am really excited about having some logo :D
18:57:31 <rluethi> yes.
18:57:40 <rluethi> let's call it a day :)
18:57:55 <julen> yes.. and they are short and dynamic, so no need to get into boring technical stuff
18:58:18 <julen> well... thanks for the meeting :)
18:58:35 <rluethi> thank you for coming.
18:58:42 <rluethi> don't forget endmeeting
18:58:45 <rluethi> :)
18:58:47 <julen> #endmeeting training-labs