17:09:03 <roshi> #startmeeting Cloud WG
17:09:03 <zodbot> Meeting started Wed Jul 29 17:09:03 2015 UTC.  The chair is roshi. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:09:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:09:03 <dustymabe> rtnpro: hey!
17:09:10 <roshi> #meetingname Cloud WG
17:09:11 <zodbot> The meeting name has been set to 'cloud_wg'
17:09:14 <rtnpro> dustymabe, what's up?
17:09:21 <roshi> #topic Roll Call
17:09:25 <roshi> who's around?
17:09:27 <rtnpro> .fas rtnpro
17:09:27 <zodbot> rtnpro: rtnpro 'Ratnadeep Debnath' <rtnpro@gmail.com>
17:09:33 <roshi> sorry for the late start :/
17:09:36 <roshi> .hello roshi
17:09:37 <zodbot> roshi: roshi 'Mike Ruckman' <mruckman@redhat.com>
17:09:41 <dustymabe> .hellomynameis dustymabe
17:09:42 <zodbot> dustymabe: dustymabe 'Dusty Mabe' <dusty@dustymabe.com>
17:09:43 <sayan> .hellomyname sayanchowdhury
17:09:52 <jbrooks> .hellomynameis jasonbrooks
17:09:53 <zodbot> jbrooks: jasonbrooks 'Jason Brooks' <JBROOKS@REDHAT.COM>
17:09:55 <kushal> .hellomynameis kushal
17:09:57 <zodbot> kushal: kushal 'Kushal Das' <kushaldas@gmail.com>
17:09:57 <sayan> .hellomynameis sayanchowdhury
17:10:09 * kushal will be on and off in the meeting.
17:10:17 <dustymabe> yay kushal is here
17:10:27 <kushal> dustymabe, :)
17:10:40 <roshi> sorry about your stomach issues kushal :(
17:10:49 <roshi> #topic Previous meeting follow-up
17:10:51 <dustymabe> are we missing anyone?
17:10:52 <roshi> * lalatenduM will setup a CI for Fedora dockerfiles
17:10:52 <roshi> * number80 will write the flyer text for Fedora Cloud flyer.
17:10:52 <roshi> * kushal to comment in https://bugzilla.redhat.com/show_bug.cgi?id=1234504 with the git commit details
17:10:55 <roshi> * kushal will file the rel-eng ticket for the updated cloud image.
17:10:57 <roshi> * dustymabe to open a ticket for default firewall behavior in Fedora Cloud images.
17:11:02 <dustymabe> looks like jzb isn't around
17:11:11 <lalatenduM> .hello lalatenduM
17:11:12 <zodbot> lalatenduM: Sorry, but you don't exist
17:11:14 <roshi> I feel like there are still cons going on
17:11:20 <lalatenduM> ahh :)
17:11:25 <lalatenduM> .hello lalatendu
17:11:28 <roshi> lalatenduM: having an existential crisis over there?
17:11:28 <zodbot> lalatenduM: lalatendu 'Lalatendu Mohanty' <lmohanty@redhat.com>
17:11:49 <dustymabe> :)
17:11:58 <kushal> So for updated cloud-images, the images are passing all tests, but the base image is throwing messages like: http://paste.fedoraproject.org/249473/18823114/
17:11:59 <lalatenduM> roshi: crisis averted :)
17:11:59 <dustymabe> so I opened the ticket for the default firewall
17:12:03 <dustymabe> #link https://fedorahosted.org/cloud/ticket/113#comment:2
17:12:22 <roshi> dustymabe: yeah, I saw some of the movement on it
17:12:27 <kushal> I need to know if I should go ahead and file a ticket to rel-eng about the updated images. Any comments?
17:12:42 <roshi> we'll have to advise the DO peeps about setting up the firewall on their host
17:13:01 <dustymabe> kushal: so I tested your networkd image and saw the same problem.. is your networkd image the same as this one or just similar becasue they are both updates?
17:13:07 <roshi> kushal: I'd file the ticket - any issues can get sussed out in the ticket
17:13:19 <kushal> dustymabe because they are updates.
17:13:24 <dustymabe> kushal: ok
17:13:33 <kushal> roshi, Okay, so keep the action item on me.
17:13:43 <kushal> I will do it today only, have it half written in a tab :)
17:13:44 <lalatenduM> regarding the CI setup , I read more doc on taskotron
17:13:50 <lalatenduM> it looks finenow
17:14:06 <roshi> #action kushal to file releng ticket for updated cloud images
17:14:18 <lalatenduM> my next plan is to install it on a local Fedora machine and do some POC
17:14:22 <dustymabe> kushal: also update our trac ticket with the status
17:14:31 <roshi> lalatenduM: a local POC of taskotron?
17:14:39 <lalatenduM> roshi: yes
17:14:45 <roshi> sweet
17:15:25 <lalatenduM> roshi: you can suggest me if u hav a better idea , as I am sure u have good experiance with taskotron
17:15:27 <roshi> if you notice things, tflink is always happy to get patches :)
17:15:29 <kushal> dustymabe, Yup.
17:15:39 <roshi> lalatenduM: we can sync up outside the meeting if you like
17:16:09 <lalatenduM> tflink: I have meet him in last qe camp , not sure if he remembers :)
17:16:23 <lalatenduM> roshi: yup, sounds good
17:16:42 <lalatenduM> s/him/you/
17:17:06 <kushal> #link https://fedorahosted.org/rel-eng/ticket/6219
17:17:10 <kushal> dustymabe, roshi ^^^
17:17:15 <roshi> thanks
17:17:51 <roshi> doesn't look like number80 is around for his update
17:18:27 <dustymabe> I'll ask in the ticket for a status update
17:18:34 <roshi> thanks
17:18:54 <roshi> kushal, can we get a tldr on the changes in that image?
17:19:00 <roshi> that way others can also test it?
17:19:20 <kushal> roshi, the networkd image?
17:19:39 <roshi> the image listed in the ticket
17:20:03 <roshi> I didn't think it was ever decided we would be for sure switching to networkd as the default stack in our images
17:20:22 <kushal> Ah, that is just with all updated packages, even I have to find out what all updated.
17:20:26 <kushal> I will update the ticket.
17:20:31 <roshi> awesome, thanks
17:20:32 <kushal> Put another action item :)
17:20:55 <roshi> #action kushal to update releng ticket #6219 with what packages were updated
17:21:29 <roshi> alright, onto the meeting items!
17:21:34 <roshi> thanks for the updates folks :)
17:21:55 <roshi> #topic Missing Cockpit RPMs (#105)
17:22:01 <roshi> #link https://fedorahosted.org/cloud/ticket/105
17:22:26 <jbrooks> I'm on this one -- I made some progress w/ a cloud-init approach, cited in the ticket
17:22:41 <jbrooks> It's not quite working right -- needs a reboot and I'm not sure why
17:23:52 <dustymabe> jbrooks: I might be able to look at this with you.. especially once monday gets here
17:24:02 <jbrooks> dustymabe, cool
17:24:11 <roshi> does that timeframe work for us?
17:24:21 * roshi doesn't see why not - just checking
17:24:37 <jbrooks> I'll look into the kickstart part of it later this week -- I sort of assume/expect that'll be easier
17:25:11 <roshi> sounds good to me
17:25:26 <roshi> can you update the ticket that this'll get looked at no later than this monday?
17:25:47 <jbrooks> Yes, I'll add a comment on that
17:26:12 <roshi> awesome, thanks :)
17:26:37 <roshi> #action jbrooks to update ticket #105 with status
17:27:08 <roshi> #action dustymabe jbrooks to work on #105 next monday (2015-08-03)
17:27:27 <roshi> #topic Investigate systemd-networkd (#14)
17:27:33 <roshi> #link https://fedorahosted.org/cloud/ticket/14
17:28:07 <roshi> as for the last comment here, I have a small issue with changing mid release
17:28:14 <roshi> mattdm: you available?
17:28:20 * dustymabe looks at latest comments
17:28:28 <roshi> I'm fine with comparing the two during F23
17:28:49 <roshi> however, swapping it at beta really defeats the purpose of have a release cycle
17:29:02 <dustymabe> mid release?
17:29:03 <roshi> we might as well not test Alpha
17:29:15 <dustymabe> roshi: are you talking about switching for F22 updated image?
17:29:22 <roshi> "...showdown for the F23 Alpha, and then decide on one for F23 Beta based on results."
17:29:46 <roshi> which I read as "compare during alpha, then change for beta if it wins the comparison."
17:29:57 <dustymabe> roshi: ahh ok you are talking about F23
17:30:00 <roshi> and I wouldn't switch it in F22 updates at all
17:30:06 <dustymabe> ++
17:30:17 <roshi> the stack should stay the same for a release - just getting updates to the currently installed tree
17:32:07 <roshi> kushal created an image to look at
17:32:18 <dustymabe> yeah I actually used it on my openstack at home
17:32:21 <roshi> but now we need to come up with some metrics/criteria to compare the two
17:32:23 <dustymabe> seemed to boot fine
17:32:45 <dustymabe> honestly I think the difference to end users is small
17:32:52 <roshi> perhaps a network load test, a suite of esoteric network configs (that you'd run into for cloud)
17:33:01 <roshi> that might be true
17:33:13 <roshi> but *all* of the cloud relies on the networking stack
17:33:29 <dustymabe> yeah.. but you can't even get into an instance if it doesn't have network
17:33:41 <dustymabe> so we're not even affecting end users
17:33:48 <dustymabe> we are affecting image builders
17:33:53 <roshi> so I think before we make a swap like this, we need to test it until it's been thoroughly tested
17:34:22 <dustymabe> roshi: what would the test criteria look like?
17:34:25 <roshi> and doing our due diligence with the comparison
17:34:29 <roshi> I'm not sure
17:34:36 <dustymabe> i'll take a stab
17:34:46 <dustymabe> 1 - Boots and network comes up on AWS
17:34:52 <dustymabe> 2 - Boots and network comes up on Openstack
17:34:53 <roshi> but I don't really see any benefit of diverging our network stack from the rest of the Editions
17:34:58 <dustymabe> 3 - Boots and network comes up on GCE
17:35:06 <dustymabe> etc..
17:35:17 <roshi> that's a good baseline
17:35:32 <roshi> here we run into another issue with a real lack of useage metrics
17:35:42 <dustymabe> the only thing we aren't accouting for is cloud-init configuration of the network
17:35:51 <roshi> we don't know how the image gets used, and what networking features people are using in the wild
17:36:03 <dustymabe> which is https://bugzilla.redhat.com/show_bug.cgi?id=1194623
17:36:24 <dustymabe> well that bug isn't an issue but the comments document an issue
17:36:47 <dustymabe> roshi: I would argue the networking features people use are low.. we only have init scripts no NM
17:36:50 <roshi> right
17:37:12 <roshi> but it's an assumption
17:37:14 <dustymabe> I would say for the most part people let the instance come up and don't configure network at all
17:37:27 <dustymabe> roshi: right, an assumption
17:38:17 <dustymabe> so if we can write down requirements
17:38:18 <roshi> I agree with gholms in comment #42
17:38:23 <dustymabe> then we can do a comparison
17:38:36 <dustymabe> and make an educated decision
17:38:40 <roshi> right, that's what I'd want to do
17:38:45 <roshi> and it's too late for F23
17:38:54 <roshi> Alpha freeze was today and we've already branched
17:39:08 <roshi> so at the earliest, this could land F24
17:39:26 <roshi> if we get the requirements written, testing done and the different groups sign off on it
17:39:31 <dustymabe> man our cycles are short
17:39:33 <dustymabe> :)
17:39:47 <roshi> fesco and QA, are the two I see (mainly), if we end up wanting a change
17:39:59 <roshi> that they are :p
17:40:30 <roshi> I'll update the ticket
17:40:37 <dustymabe> roshi: can you add to the ticket a request to have people add requirements and maybe link to a page where this information can live
17:41:12 <roshi> specifically with: we need to write requirements for our networking stack, that it's too late for F23, and that we'll need input from other groups
17:41:16 <roshi> yeah
17:41:31 <roshi> I'll start a wiki page in the cloud space for discussion
17:41:47 <dustymabe> roshi: thanks!
17:42:03 <roshi> np
17:42:16 <roshi> #action roshi to update ticket #14 with status
17:42:36 <roshi> #topic Care and feeding of Dockerfiles
17:42:42 <roshi> #link https://fedorahosted.org/cloud/ticket/84
17:43:08 <roshi> eh, we can skip this - still need to break out into individual tickets
17:43:33 <roshi> I don't have the cycles right now to break these out
17:43:36 <roshi> does anyone else?
17:43:48 <roshi> we're running short on time ... :/
17:43:51 <lalatenduM> I have given a suggestion wrt this
17:44:05 <lalatenduM> i.e. fedora dockerfiles
17:44:22 <lalatenduM> we should maintain different git branches
17:44:27 <lalatenduM> for fedora releases
17:44:42 <dustymabe> lalatenduM: did you propose this on the list already?
17:44:52 <lalatenduM> dustymabe: yes
17:44:53 <dustymabe> I'm +1 for doing this
17:45:14 <roshi> that makes sense
17:45:18 <dustymabe> I say make it happen
17:45:20 <lalatenduM> how to move ahead as it seems everyone liked the idea
17:45:29 <roshi> lalatenduM: can you update the ticket with a link to your post to the list?
17:45:38 <lalatenduM> roshi: sure
17:46:06 <dustymabe> lalatenduM: I'd say best strategy is try to get all of them to work good for F22
17:46:11 <roshi> #action lalatenduM to update ticket #84 with his post to the mailing list
17:46:16 <dustymabe> and then make F22 branch
17:46:21 <lalatenduM> dustymabe: yup
17:46:27 <dustymabe> and then we can make new branches as we go
17:46:28 <roshi> or make a F22 branch, and then make them work
17:46:29 <lalatenduM> and for rawhide
17:46:35 <roshi> what do they work mostly on now?
17:46:42 <dustymabe> roshi: it's a mix
17:46:49 <roshi> iirc that a lot of them still required yum
17:46:54 <dustymabe> depending on who had an itch to scratch
17:46:54 <lalatenduM> roshi: we have just master branch
17:47:02 <roshi> or needed to be ported to use dnf, rather
17:47:24 <lalatenduM> roshi: yes
17:47:50 <dustymabe> lalatenduM: if you could update the ticket or the list thread with some of this strategy it might help move the discussion
17:47:59 <roshi> thanks lalatenduM :)
17:48:07 <lalatenduM> dustymabe: the mail has startegy in it
17:48:17 <roshi> we're running out of time, going to move on if that's ok
17:48:28 <roshi> gonna skip #108 since jzb isn't here
17:48:39 <roshi> (#108 is text for the cloud flyer)
17:49:57 <dustymabe> ok
17:50:02 <dustymabe> open floor?
17:50:17 <roshi> sure
17:50:23 <roshi> #topic Open Floor
17:50:54 <roshi> https://fedorahosted.org/cloud/ticket/107 we didn't get much discussion on this, but I posted it to the council trac
17:51:23 <roshi> https://fedorahosted.org/council/ticket/36
17:51:51 <dustymabe> I don't have anything for open floor this week
17:52:02 <roshi> and dustymabe I'm starting a convo with the DO peeps about what they'd need if we had to get them to have the firewall installed/configured going forward
17:54:02 <roshi> which I guess is just a PR to your repo on GH :p
17:54:27 <roshi> anyone have anything else?
17:55:31 <roshi> since we ran short on time, this week go through the tickets [0] and leave comments as you see fit
17:55:31 <dustymabe> roshi: set the fuse
17:55:35 <roshi> [0] https://fedorahosted.org/cloud/report/9
17:55:41 * roshi sets the fuse
17:55:45 <roshi> 3...
17:56:13 <roshi> 2...
17:56:15 <dustymabe> as for DO yeah let's just talk to them in the slack channel. I should be able to update the script in the repo and we can use it for the updated image that we release
17:56:30 <roshi> sounds like a plan
17:56:36 * roshi can help with that
17:56:43 <roshi> 1...
17:56:48 <roshi> thanks for coming folks!
17:57:03 <roshi> #endmeeting