14:32:16 <tflink> #startmeeting Fedora QA meeting
14:32:16 <zodbot> Meeting started Mon Aug  5 14:32:16 2013 UTC.  The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:32:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:32:19 <tflink> #meetingname fedora-qa
14:32:19 <zodbot> The meeting name has been set to 'fedora-qa'
14:32:24 <tflink> #topic Roll Call
14:32:34 <tflink> #chair adamw kparal
14:32:34 <zodbot> Current chairs: adamw kparal tflink
14:32:51 <tflink> anyone here to be called a roll?
14:32:56 * satellit listening
14:33:23 * mkrizek is here
14:33:39 * atodorov is here (will introduce myself later)
14:33:49 * kparal joins late
14:33:52 * spoore is listening
14:34:31 * pschindl is here
14:34:56 * tflink waits a few more minutes as more people join
14:37:23 <tflink> ok, let's get started
14:37:39 * tflink is ad-libbing a bit, hopefully will cover the parts that adamw had in mind
14:38:08 <tflink> #topic Fedora 20 Planning/Change Review
14:38:25 * handsome_pirate is here, btw
14:39:18 <tflink> supposedly TC1 is scheduled for this week
14:39:24 <tflink> but I can't find the date
14:39:52 <tflink> oh, I think this schedule has been changed recently
14:40:37 <tflink> yep, everything got shifted back by ~ 2 weeks
14:41:15 <tflink> #info fedora 20 branch is now scheduled for "no later than 2013-08-20"
14:41:22 <nirik> yes, it's moved out 2 weeks from the 'no later than' thing.
14:42:01 <tflink> #info fedora 20 alpha TC1 is now scheduled for 2013-08-20
14:42:17 <tflink> #link http://fedoraproject.org/wiki/Releases/20/Schedule
14:42:33 <tflink> #link http://fedorapeople.org/groups/schedule/f-20/f-20-quality-tasks.html
14:43:50 * tflink is reading through the changes marked in the meeting outline
14:44:21 <nirik> I think that was decided as the actual schedule... but perhaps it didn't get enough votes.
14:44:55 <tflink> nirik: do you know if that was in a fesco ticket or meeting minutes?
14:45:04 <nirik> in ticket, let me look
14:45:13 <handsome_pirate> Either way, I'm +1 for more time
14:45:15 <nirik> https://fedorahosted.org/fesco/ticket/1095#comment:22
14:45:42 <nirik> so if folks don't think thats reasonable, do bring it up there. Otherwise the +2 week version is what we have
14:47:09 <tflink> I'm sure there will be lots of happy us-ians about thanksgiving plans :)
14:47:46 <tflink> nirik: we were planning on the old schedule, so another 2 weeks isn't going to be a horrible thing :)
14:48:03 <nirik> yeah
14:48:16 * handsome_pirate will survive working through Thanksgiving
14:48:18 <nirik> ditto on both accounts.
14:48:58 <tflink> the timing may get interesting for other reasons, but let's cross that bridge if/when we get there
14:49:16 <tflink> so, the list of features identified for attention today are:
14:49:44 <tflink> System Wide:
14:49:44 <tflink> LabeledNFS
14:49:44 <tflink> NetworkManagerBonding
14:49:45 <tflink> NetworkManagerBridging
14:49:45 <tflink> VisibleCloud
14:49:47 * handsome_pirate just cced himself to schedule ticket
14:49:57 <tflink> Self Contained:
14:49:57 <tflink> NetworkManagerCLIAddConnection
14:50:18 <tflink> Links available from wiki page: https://fedoraproject.org/wiki/QA/Meetings/20130805
14:50:31 <tflink> anything in particular that people feel should be discussed?
14:51:33 * tflink interprets silence as a "not so much"
14:51:56 * handsome_pirate is looking through them
14:52:08 <tflink> I think adam had concerns about visable cloud and our lack of cloudy testing but I don't know more details than that
14:52:26 <handsome_pirate> LabeledNFS may want some attention
14:52:28 <tflink> s/visable/visible
14:52:46 <tflink> handsome_pirate: yeah, that could cause issues with nfs sourced install media
14:52:55 <tflink> in addition to any more testing
14:53:04 * tflink is looking forward to bridging support in NM
14:53:38 <handsome_pirate> +1 bridging
14:53:52 * atodorov has an old RFE open for VPN over NM bonding
14:53:53 <handsome_pirate> The question that keeps coming up with cloud is how do we test?
14:54:05 <handsome_pirate> re:  visible cloud
14:54:09 <tflink> handsome_pirate: not sure, honestly
14:54:30 <handsome_pirate> For instance, I don't have a credit card, so I can't spin up an Amazon instance to test it
14:54:33 <tflink> I think that the expectation is that people are either willing to pay for EC2 or have access to some openstack thing
14:54:52 <handsome_pirate> Ah
14:54:57 <atodorov> handsome_pirate: I have RedHat backed EC2 account, also specialize in installation testing
14:55:12 <handsome_pirate> atodorov:  But, I don't
14:55:16 <handsome_pirate> I don't work for RH
14:56:17 <tflink> handsome_pirate: I really don't know what the plan is there - the cost barrier to EC2 testing is one issue, though
14:56:30 <atodorov> handsome_pirate, tflink: reasouce issue aside, what do you want  to test for cloud images on amazon ?
14:56:38 <kparal> the cloud people tested it themselves in the past releases, I don't expect it to change this time
14:57:03 <Southern_Gentlem> kparal,  but better to communicate than assume
14:57:21 <handsome_pirate> kparal:  Well, if it's going to be an official (almost primary) release, should QA not give it a go?
14:57:22 <tflink> kparal: I think the concern is that the cloud folks really aren't testing all that much and if we want to be advertising more, qa should be testing in some form
14:57:44 <handsome_pirate> ^^
14:57:45 <tflink> atodorov: that's a good question, really. cloud isn't all that different from VMs other than initial boot
14:58:26 * ignatenkobrain here
14:58:29 <atodorov> tflink: agree, if it boots then it works :). for EC2 however we may want to test some of their AMI creation guidelines, like for example remove your ssh keys from the image
14:58:30 <tflink> there's a lot of "OMG, it's cloud and therefore all complicated and stuff ..." which it is, but not as much from the users perspective
14:58:31 <ignatenkobrain> hi all !
14:59:00 * handsome_pirate gets that
14:59:07 <handsome_pirate> It's a glorified VM
14:59:11 <tflink> atodorov: good point, especially with the concerns raised about digitalocean as of late
14:59:27 <handsome_pirate> I just think that QA should pass off on it if we're going to be making it as visible as it's going to be
14:59:38 <atodorov> tflink: I'm not aware of this. Can you send me a link pls
15:00:18 <tflink> atodorov: http://missingm.co/2013/07/identical-droplets-in-the-digitalocean-regenerate-your-ubuntu-ssh-host-keys-now/
15:01:01 <tflink> handsome_pirate: sure, but now we've come full circle to "what do we want/need to test about cloud images?"
15:01:38 <handsome_pirate> tflink:  We don't exactly have much in the way of criteria, do we?
15:01:41 <tflink> I'd love to test more of the qcow2 images, I *think* they're pretty much the same as AMIs content-wise
15:01:51 <handsome_pirate> tflink:  I assume that at least the minimal criteria would apply
15:01:56 <tflink> handsome_pirate: because they're pretty much glorified VMs
15:02:05 <handsome_pirate> tflink:  Aye
15:02:11 <handsome_pirate> As I said above
15:02:20 <tflink> the fancy stuff is more os/distro agnosti
15:02:23 <tflink> agnostic
15:02:40 <handsome_pirate> It would be cool to make sure it works
15:02:43 <tflink> handsome_pirate: yep, I liked the phrase so I'm stealing it :-P
15:02:47 <handsome_pirate> heh
15:02:58 <handsome_pirate> It's Creative Commons :)
15:03:12 <tflink> anyhow, we're getting off into the weeds a bit
15:03:16 <handsome_pirate> Aye
15:03:41 <atodorov> #link http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AESDG-chapter-sharingamis.html#public-amis-identity
15:03:47 <tflink> #info we're still not clear on what exactly should be tested for cloud images other than "do they boot"
15:04:19 <tflink> #info atodorov suggested that the EC2 AMIs could be checked for amazon's AMI creation guidelines
15:04:44 <tflink> is ^^^
15:04:50 <tflink> phrased well enough to understand?
15:05:11 <handsome_pirate> Aye
15:05:17 <atodorov> yup
15:05:31 <tflink> I wonder if that fits better into the automation stuff that the cloud SIG has been talking about
15:05:45 * tflink hasn't asked them about their progress on that front lately
15:05:54 <handsome_pirate> tflink:  I can ping them about it
15:06:35 <tflink> handsome_pirate: otherwise I'm sure it'll come up @ flock
15:06:51 <handsome_pirate> True
15:07:31 <tflink> #info there  are also concerns about how to test cloud images w/o an EC2 account and whether we want/need to worry about granting access to either an account for testing or an openstack instance - will require more discussion
15:07:41 <tflink> OK, any more concerns about cloud
15:07:42 <tflink> ?
15:07:51 * adamw apparently can't read clocks
15:08:05 <adamw> morning folks
15:08:10 * tflink has always been of the opinion that clocks are overrated :)
15:08:26 <tflink> adamw: you've already been chaired, feel free to take over
15:08:40 * tflink isn't sure if there were specific concerns you had about the listed features
15:08:43 * handsome_pirate thinks adamw should fire himself
15:08:54 <adamw> i'm with handsome_pirate on the visible cloud stuff, that's why i put that Change on the meeting list, fwiw
15:09:03 <adamw> the others are just changes that are new since last meeting
15:09:24 <adamw> tflink: you may as well keep driving
15:09:27 <tflink> tagged nfs might be interesting for install testing
15:09:40 <tflink> adamw: ok
15:09:56 <tflink> any other concerns about visible cloud images?
15:10:24 * satellit afk
15:10:42 <tflink> #topic Fedora 20 Planning/Change Review - LabeledNFS
15:10:46 <adamw> i haven't read all the scrollback, but my thinking was we might need to tweak the criteria a bit and maybe add a test case or two, just for the t hings that have been discussed
15:10:49 * nirik notes we could look at granting some qa folks access to the infra private openstack cloud...
15:10:59 <nirik> (if testing in openstack is desired)
15:11:03 <adamw> check the image boots and any important 'first time steps' for cloud images fly
15:11:22 <tflink> nirik: the thought was that it might be a better option than EC2 testing most of the time
15:11:28 <tflink> it -> openstack
15:11:51 <handsome_pirate> +1
15:11:58 <nirik> yeah, could be.
15:12:16 <nirik> but they are quite different too... so both would be best.
15:13:23 <tflink> nirik: true but I figured that we could get the common stuff done with openstack to reduce the amount of paid testing needed for EC2
15:13:38 * nirik nods
15:13:57 <atodorov> tflink: we can use some spot instances in EC2 as well to reduce the cost further
15:13:57 <tflink> #undo
15:13:57 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x249ba610>
15:14:14 <tflink> atodorov: true, I've certainly done that before and saved money
15:14:35 <tflink> as long as you don't mind getting an instance ripped out from underneath you suddenly :)
15:14:43 <tflink> which has happened to me
15:14:55 <tflink> and others, I assume
15:15:20 <atodorov> tflink: for personal use I set a high enough spot bid and haven't seen an instance shutting down on me
15:15:24 <tflink> #info testing cloud images on infra's private cloud may be an option for some testing
15:16:08 <tflink> atodorov: yeah, I learned about not setting the bid too low when my hadoop cluster fell over when instances got yanked :)
15:16:28 <tflink> ok, anything else before moving on? we have ~ 15 minutes left
15:16:35 * handsome_pirate is good
15:16:54 <tflink> #topic Fedora 20 Planning/Change Review - LabeledNFS
15:17:12 <tflink> I should have split off the cloud stuff as well but either way
15:17:41 <tflink> my main question here is whether we want to account for labeled nfs for install tesitng
15:17:44 <tflink> testing
15:17:52 * tflink apparently can't spell today
15:18:08 <handsome_pirate> We should
15:18:17 <adamw> how do you mean 'account for'?
15:18:29 <tflink> adamw: testing installs with and without
15:18:44 <adamw> so basically run the NFS install tests against F20 servers?
15:18:45 <tflink> not sure it'll make a huge difference, though. anaconda runs in permissive
15:19:12 <tflink> adamw: it sounds like less of a concern when you put it that way
15:19:44 <tflink> so probably not needing to call out install tests with labeled nfs so much
15:20:01 <handsome_pirate> Still, it may not be a bad idea to do it once, just to check
15:20:12 <adamw> sure
15:20:42 <atodorov> tflink, adamw: anaconda runs in permissive but who handles access enforcement - client or server ?
15:20:54 <tflink> atodorov: good question
15:21:00 <tflink> server, I assue
15:21:11 <tflink> NFSv3 is RPC based
15:21:16 <atodorov> then it is possible to get a read error
15:21:26 <tflink> I assume that NFSv4 is going to be similar that way
15:22:03 <tflink> but I've never really gotten into how NFSv4 actually works
15:22:04 <handsome_pirate> But, assumptions ...
15:23:06 <tflink> yeah, I know
15:23:35 <tflink> #info we should probably test nfs sourced installs from a F20 server with labeled nfs some
15:23:40 <tflink> #undo
15:23:40 <zodbot> Removing item from minutes: <MeetBot.items.Info object at 0x26085f90>
15:24:00 <tflink> #info we should probably test nfs sourced installs from a F20 server with labeled nfs some - it's not clear whether there could be read errors on the server even though anaconda runs in permissive
15:24:34 <tflink> any other stuff to bring up before open floor?
15:25:01 <tflink> atodorov: do you want to do a bit of an introduction?
15:25:41 <atodorov> tflink: yes, I will be quick
15:25:51 <tflink> #chair atodorov
15:25:51 <zodbot> Current chairs: adamw atodorov kparal tflink
15:26:09 <tflink> atodorov: go ahead, not sure what you want to call the topic
15:26:35 <kparal> #topic Fedora QA and RTT cooperation
15:26:46 <tflink> kparal: that works :)
15:26:54 <atodorov> #info atodorov is joining (part time) Fedora QA and helping with installation related items short term. Long term I'd like to work on ifrastructure too. I have 6+ years of anacodna testing experience for RHEL
15:27:06 <atodorov> that's the short info ^^^
15:27:19 <adamw> welcome!
15:27:24 <handsome_pirate> Ahoy!
15:27:27 <atodorov> I'm coming from RedHat's internal team that work on installation testing for RHEL
15:27:28 <tflink> welcome to the party!
15:27:32 <kparal> atodorov should be running some RTT tests in house and sending us results on the TCs and RCs
15:27:52 <kparal> that could save us a ton of manual work, or at least identify some interesting bugs
15:27:59 <atodorov> I will be working with kamil in the short term and bring some of the internal installation testing wisdom and automation to Fedora
15:28:03 <tflink> I forget, what is RTT short for?
15:28:05 * handsome_pirate is happy to see more vict... I mean helpers
15:28:08 <kparal> Release Test Team
15:28:14 <atodorov> RTT == Release Test Team
15:28:23 <atodorov> aka Installation QE
15:28:25 <tflink> thanks, I knew what they did, just not what RTT meant :)
15:29:06 <tflink> atodorov, kparal: anything else you guys want to cover before moving into open floor?
15:29:21 <atodorov> one more thing
15:29:41 <atodorov> long term I'd like to see a Fedora QA lab (infrastructure) like we have for RHEL - all automated
15:29:53 <atodorov> feel free to include me in meetings and such on this topic
15:30:01 <tflink> atodorov: we're working on it :)
15:30:04 <atodorov> I can do Python and Django and Bash coding ;)
15:30:21 <kparal> ok, thanks
15:30:30 <handsome_pirate> atodorov:  See tflink's taskbot blog posts
15:30:48 <atodorov> handsome_pirate: will do, still getting up to speed
15:31:27 * handsome_pirate has to go
15:31:33 <handsome_pirate> peace, y;all
15:31:38 <tflink> atodorov: I'll have some questions about how RTT automates installs but that'll be for another time
15:31:49 * atodorov has worked on Beaker patches for EC2 in the past but they never got merged back I think
15:32:12 <atodorov> tflink: I will introduce myself on the mailing list and you can ping me via email
15:32:17 <tflink> anyhow, we're over time so if there's nothing else that needs to be covered in meeting ...
15:32:21 <tflink> atodorov: sounds like a plan
15:33:05 <tflink> #topic Open Floor
15:33:11 <tflink> Any items for open floor?
15:34:27 * tflink sets the Fedora QA patent-pending non-deterministic fuse for [0,5] minutes
15:35:47 <tflink> boom!
15:35:54 <tflink> Thanks for coming everyone!
15:35:57 <tflink> #endmeeting