14:02:08 <uli-k> #startmeeting Octopus weekly meeting
14:02:08 <collabot> Meeting started Mon Jul 27 14:02:08 2015 UTC.  The chair is uli-k. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:08 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:02:08 <collabot> The meeting name has been set to 'octopus_weekly_meeting'
14:02:17 <uli-k> Hi everybody!
14:02:28 <ran> hello
14:02:33 <meimei1> hi,Uli
14:02:43 <narindergupta> #info Narinder Gupta
14:02:54 <meimei1> #info meimei
14:03:03 <uli-k> #topic roll call
14:03:11 <uli-k> #info Uli
14:04:50 <uli-k> looks like we are a small group today. Is everybody in vacation or on ODL summit?
14:05:19 <meimei1> Maybe
14:05:49 <icbts> ODSummit
14:05:52 <uli-k> anyways... We should have a short meeting.
14:06:15 <meimei1> is there a  agenda? Uli
14:06:35 <uli-k> I put some standard agenda on the wiki...
14:06:43 <uli-k> #topic Agenda Bashing
14:07:37 <uli-k> #link https://wiki.opnfv.org/meetings/octopus#agenda_of_next_meeting_monday_july_13_1400_utc ---- forgot to change the date. Same agenda as before, since we have same topics.
14:08:10 <fdegir> #info Fatih Degirmenci
14:08:35 <uli-k> Hey fdegir, good to "see" you.
14:08:38 * fdegir has limited connectivity - just chat, but would like to mention some stuff towards the end
14:08:48 <fdegir> hi everyone
14:09:15 <uli-k> hi
14:09:33 <edavis13> hi Uli
14:09:40 <edavis13> <--- Broadcom  :-)
14:09:54 <uli-k> We had some good internal discussion on CI pipeline last week.
14:10:04 <uli-k> But let's go for the agenda first
14:10:19 <uli-k> any additions (except fdegir "at the end"...)
14:10:29 <uli-k> ?
14:10:45 <uli-k> OK. Then let's go Action items.
14:10:54 <uli-k> #topic action items review
14:11:50 <uli-k> chigang wanted to work on naming scheme for slaves connected to Jenkins
14:12:17 <uli-k> he didn't info, so let's keep it.
14:12:33 <uli-k> #action chigang to propose naming scheme for slaves connected to Jenkins on https://etherpad.opnfv.org/p/octopusR2
14:14:09 <uli-k> fdegir, I see an action "fdegir to add this to etherpad "
14:14:23 <uli-k> Do you remember what to put?
14:14:27 <fdegir> yes
14:15:09 <uli-k> It was on: frankbrockners reminds that we have to prirotize the work needed to do the sequential execution on a single POD - build/deploy/test for multiple installers on one node
14:15:09 <fdegir> #info frankbrockners reminded us that we should start working on sequential runs in CI since we're reconfiguring LF lab
14:15:17 <uli-k> :D
14:15:32 <fdegir> :)
14:16:00 <fdegir> #info the information has already been added to etherpad and pbandzi, trozet, and aricg started working on it already
14:16:16 <uli-k> Do you have Jira numbers?
14:16:37 <fdegir> looking at it now
14:17:12 <pbandzi> #info JIRA: OCTO-109
14:17:22 <fdegir> thx pbandzi
14:17:23 <meimei1> fatih, I have a question, why not use other labs?
14:17:43 <meimei1> any lab connect to jenkins master
14:17:55 <fdegir> meimei1: pharos is yet to come up with lab conformance spec
14:18:15 <fdegir> and the lab owners are yet ro come up with lab specs
14:18:33 <fdegir> so we have labs but we don't have enough info about them
14:18:42 <fdegir> or no requirments on them
14:18:56 <meimei1> what can we do about that?
14:19:04 <fdegir> this work is also ongoing and driven by trevor/pharos
14:19:38 <uli-k> Do you know whether Trevor needs help?
14:20:12 <fdegir> uli-k: I think so
14:20:29 <uli-k> Meimei, can I action you?
14:20:30 <fdegir> we can at least help with reviewing
14:20:58 <fdegir> the thing is; we need 2 templates: one for installer requirements
14:20:59 <meimei1> yes
14:21:09 <fdegir> and the other is for lab info
14:21:18 <uli-k> #action meimei to contact Pharos and find some help to drive conformance spec.
14:21:29 <fdegir> once these templates are ready, installer projects and lab owners need to complete those documents
14:21:30 <meimei1> ok
14:21:44 <uli-k> #undo
14:21:44 <collabot> Removing item from minutes: <MeetBot.ircmeeting.items.Action object at 0x249cf50>
14:21:51 <fdegir> and the conformance is the 3rd document
14:21:59 <uli-k> #action meimei1 to contact Pharos and find some help to drive conformance spec.
14:22:36 <fdegir> in short, we have to make sure we are able to run CI with only 1 POD
14:23:07 <fdegir> and even though all the labs are functional, LF POD(s) might be the place where the release are going out
14:23:20 <uli-k> yes. And later, when we have multiple labs, we could do round robin on a few of the labs.
14:23:30 <fdegir> yes
14:23:35 <uli-k> labs or PODs
14:23:42 <uli-k> OK. next.
14:23:45 <meimei1> and  then we can utilize all those env?
14:23:49 <fdegir> we as Octopus should also have some kind of acceptance criteria on community labs
14:24:51 <uli-k> You mean additional to Pharos criteria?
14:25:06 <fdegir> uli-k: I think so
14:25:09 <uli-k> Can't we add our reqs to Pharos?
14:25:25 <fdegir> the requirements should go to pharos
14:25:42 <meimei1> I know that fatih mean, octopus must have some criteria
14:25:43 <fdegir> but once lab owners claim they are pharos compliant
14:26:01 <fdegir> we should run some stuff on them to see we get same kind of result from them
14:26:30 <fdegir> we should perhaps sign off the compliance by doing this
14:26:51 <fdegir> labs go to pharos saying they are ready
14:26:58 <fdegir> pharos announces this some way
14:27:04 <fdegir> and octopus validates it
14:27:14 <uli-k> ok. understand.
14:27:15 <fdegir> and accepts it as working lab/whatever
14:28:01 <fdegir> so my Action Item is done
14:28:08 <uli-k> yes :D
14:28:47 <uli-k> #agree when labs claim to be Pharos compliant, octopus will validate that.
14:28:58 <meimei1> and then we can run the jenkins jobs on those lab?
14:29:03 <uli-k> yes.
14:29:33 <uli-k> Then we can round robin for CI-Pipeline or create some intelligent assignment of tests to PODs.
14:29:39 <fdegir> pharos will have some kind of booking system
14:29:57 <fdegir> and we will see all the labs/pods as a pool of resources
14:30:07 <meimei1> on wiki?
14:30:28 <fdegir> at the start, it could/would be a simple text file or something
14:30:37 <fdegir> later on more intelligent booking system
14:31:00 <meimei1> exactly
14:31:00 <fdegir> but it's far more later
14:31:04 <uli-k> I think in the end, they need a real tool for that. In wiki anybody can change the booking...
14:31:13 <fdegir> uli-k: yes
14:31:39 <uli-k> Let's move to the next AI.
14:31:56 <uli-k> There is one more for fdegir, and probably you have done it :D
14:32:05 <uli-k> fdegir to contact LF and Pharos regarding LF Lab reconfiguration - assigning one of the PODs for CI
14:32:11 <fdegir> yep
14:32:41 <fdegir> #info pbandzi, trozet, and aricg are working on the script to do auto configuration
14:32:59 <fdegir> #info OCTO-109
14:33:37 <uli-k> #info Good. So we close the AI as well
14:33:41 <fdegir> #info until the script is ready - meaning that we have ability to automatically configure the POD - we won't touch the PODs
14:34:33 <fdegir> pbandzi: can you give a quick status update regarding the script work pleaße?
14:34:43 <pbandzi> yes
14:35:36 <fdegir> pbandzi: when do you think it will be ready?
14:35:42 <pbandzi> #info script is now in repository we had some issue and put couple of fixes to repo (for sreconfiguration and cript and jjb script)
14:36:13 <pbandzi> #info now we are having issue with jenkins verify job
14:36:27 <pbandzi> #info maight be cause by python version - need to verify
14:36:57 <pbandzi> #info expected to be working during this week
14:36:59 <pbandzi> hopefully
14:37:00 <pbandzi> end
14:37:08 <fdegir> thx pbandzi
14:37:42 <fdegir> #info once the script is ready, fuel and foreman need to be verified on  single LF POD
14:37:53 <fdegir> #info due to Arno SR1 work
14:38:17 <meimei1> only these two?
14:38:37 <fdegir> meimei1: to start with since we have info about these installers
14:39:03 <fdegir> once they're done then other installers need to provide info including what type of network confog they need, etc
14:39:18 <meimei1> ok, I got it
14:39:19 <fdegir> and the pbandzi's script configuration needs to be updated
14:39:31 <fdegir> for the rest of the installers
14:39:57 <uli-k> and probably we should test a lot of different sequence (will happen anyways.).
14:39:59 <fdegir> meimei1: perhaps tou can take a look what is there for fuel/foreman and prepare the input for compass already kow
14:39:59 <fdegir> now
14:40:34 <fdegir> uli-k: yes,there are many different combinations of things
14:40:41 <uli-k> Should we plan for a document what an installer needs to provide?
14:40:53 <meimei1> ok
14:41:09 <fdegir> uli-k: that's pharos template I mentioned above
14:41:27 <fdegir> they will use foreman/fuel docs as input and come up with template
14:41:41 <fdegir> saying that installer should provide this, this, and that info
14:42:05 <uli-k> ok. get it. So the template is a template for the installers for POD preparation.
14:42:06 <fdegir> and then installer projects will use the template and provide the info
14:42:26 <uli-k> Will the script directly read the template?
14:42:26 <fdegir> uli-k: yes; 1 for installers and - for labs
14:42:45 <meimei1> can we say that, if all the labs can be utilized, resource problem is solved?
14:43:00 <fdegir> I think the information needs to be refelcted to pbadnzi's script's configuration file
14:43:06 <uli-k> meimei1, guess not.
14:43:17 <uli-k> If there are many tests, we need many PODs.
14:43:23 <fdegir> agree with uli-k
14:43:40 <fdegir> the community labs are ther for development/developer works as qell
14:43:51 <uli-k> But if we can use all the PODs of all labs, we will be much better than today.
14:44:57 <fdegir> that was the very quick update regarding what's been happening with lab, scripts etc.
14:45:04 <uli-k> This looks like good progress, and we see where we will go.
14:45:07 <uli-k> Next AI.
14:45:33 <uli-k> MatthewLi to demo Zuul on 27th of July.
14:45:43 <uli-k> MatthewLi, you there?
14:46:03 <chigang> MatthewLi is offline
14:46:09 <uli-k> Sorry, I didn't check before setting agenda.
14:46:25 <chigang> becase his wifi does not work
14:46:34 <fdegir> I think he already recorded the demo and put it to youtube
14:46:37 <fdegir> https://www.youtube.com/watch?v=fOYdsNdkCuw
14:46:58 <meimei1> yes, that's his demo
14:47:10 <fdegir> #info Matthew have made a demo of the CI system which I am using in my lab, zuul, jenkins, logserver are used there
14:47:26 <fdegir> #info  maybe many people are on their vocation, so I give the link here first, https://www.youtube.com/watch?v=fOYdsNdkCuw
14:47:38 <uli-k> OK then we don't need to spend a weekly meeting.
14:47:42 * fdegir copied from #opnfv-octopus channel
14:48:06 <fdegir> #info any questions, please contact MatthewLi
14:48:22 <uli-k> We just need to socialize with community.
14:48:52 <uli-k> AI closed.
14:49:08 <uli-k> we skipped one before.
14:49:31 <uli-k> Action chigang to propose naming scheme for slaves connected to Jenkins
14:49:59 <chigang> uli-k: OK.
14:50:44 <chigang> uli-k: I have write it on Etherpad. please review it and give me some suggestion
14:50:56 <uli-k> R2 etherpad?
14:51:03 <chigang> uli-k: Yes.
14:51:32 <uli-k> OK.
14:51:44 <fdegir> just a reminder regarding this AI
14:51:47 <uli-k> #info everybody please review https://etherpad.opnfv.org/p/octopusR2 line 35.
14:52:11 <fdegir> #info we need to apply new naming scheme to one slave to see if it requires change of token
14:52:42 <fdegir> #info if it does, the naming scheme needs to be applied while the lab owners are online or we lose the slave until owners are back
14:52:54 <chigang> fdegir: yep, we can try it in a POD
14:54:08 <uli-k> ok. Let's do a new action for that.
14:54:45 <uli-k> #action chigang to try on a POD to apply new naming scheme to a slave.
14:55:03 <uli-k> ok. we are done with action items
14:55:09 <fdegir> :)
14:55:16 <uli-k> but now time is nearly up.
14:55:52 <uli-k> I think we covered some important issues, so I don't feel bad :D
14:55:57 <fdegir> uli-k: can you give a quick info regarding your internal octopus discussions?
14:56:06 <fdegir> 1 or 2 sentences perhaps
14:57:04 <uli-k> yes. We mainly discussed how CI pipeline can run with multiple labs (PODs to be exact).
14:58:03 <uli-k> There are a few steps after we did what we discussed today with sequential jobs on a single POD.
14:58:35 <fdegir> ok so we're moving towards the same target
14:58:38 <uli-k> First, we can round-robin a few PODs as soon as necessary.
14:58:51 <uli-k> Then we have to do more intelligent things.
14:59:14 <fdegir> yep
14:59:15 <uli-k> There might be a patch that needs to be verified on different hardware types, that is in different labs.
14:59:58 <uli-k> So we need a mechanism, to mark a patch to be run through pipeline twice or on a specific hardware
15:00:18 <uli-k> same applies on our testing.
15:00:59 <uli-k> functest or yardstick will issue test to be queued.
15:01:20 <uli-k> So similar lab selection as for normal CI pipeline will go for tests.
15:01:29 <frankbrockners> Hi folks... when do you plan to be done?
15:01:48 <uli-k> just answering last question.
15:02:02 <uli-k> let's move into octopus channel for discussion.
15:02:12 <uli-k> thanks everybody for participating.
15:02:16 <frankbrockners> BGS team likes that :-)
15:02:19 <uli-k> #endmeeting