07:59:32 <eranrom> #startmeeting storlets
07:59:33 <openstack> Meeting started Tue Sep 27 07:59:32 2016 UTC and is due to finish in 60 minutes.  The chair is eranrom. Information about MeetBot at http://wiki.debian.org/MeetBot.
07:59:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
07:59:36 <openstack> The meeting name has been set to 'storlets'
07:59:43 <eranrom> Hi
08:00:17 <takashi> eranrom: Hi
08:00:48 <eranrom> takashi: Hi, do you know if we should wait for Kota?
08:01:59 <takashi> eranrom: I didn't hear his plan.
08:02:13 <eranrom> takashi: ok, so lets start
08:02:18 <eranrom> #topic Design Summit planning
08:02:41 <eranrom> I have started to populate the ehterpad: https://etherpad.openstack.org/p/storlets-otaca-design-summit
08:03:19 <eranrom> Other then adding topics, we should probably decide on the best usage of the fb session we seem to be given.
08:03:41 <eranrom> or might be given...
08:04:17 <takashi> eranrom: ok
08:05:18 <takashi> eranrom: Can I ask the suggeed time frame for Thursday?
08:05:48 <eranrom> takashi: All I know is that its in the morning and conflicting with Swift.
08:06:04 <eranrom> We could try and see the exact time where Swift has a fb session in the morning.
08:06:27 <takashi> eranrom: ok. will check that.
08:06:36 <eranrom> takashi: Thanks!
08:07:22 <eranrom> Regarding the topic, I wanted to cover spark-storlets and you had an idea to get deeper into storlets
08:07:49 <takashi> I'll add some topic to the etherpad. sorry to be late for that.
08:08:50 <eranrom> takashi: sure, no problem at all! We can continue from there, keeping in mind that we may want to split the session between subjects...
08:09:10 <eranrom> that is for the fb session
08:09:27 <takashi> eranrom: makes sense
08:09:42 <eranrom> takashi: ok. AFAIK, Kota will not be on Fri right?
08:09:55 <takashi> I think we need to ask to kota_ and Doron (and others?) to decide when we have the design summit session,
08:10:03 <takashi> eranrom: yes, AFAIK, too
08:10:28 <eranrom> takashi: yep, I guess that for the working sessions we should not have a space problem :-)
08:10:40 <eranrom> BTW Doron will not attend thesummit
08:10:47 <takashi> eranrom: ok
08:10:57 <eranrom> anything else on this?
08:11:13 <takashi> eranrom: I'm currently checking my schedule because there are some other sessions I need to attend
08:11:22 <eranrom> takashi: sure.
08:11:26 <takashi> I'll let you know if I get any updates about that
08:11:32 <eranrom> Thanks\
08:11:47 <eranrom> #topic Big tent updates
08:13:03 <takashi> I'd like to let you know the latest status about package work about python storlet work.
08:13:21 * eranrom is all ears
08:13:33 <takashi> :-)
08:13:39 <takashi> so first, about package work
08:14:18 <takashi> Last week I finally I passed gate functional test about my patch 370332
08:14:34 <eranrom> takashi: Kudos!
08:14:39 <takashi> (Currently I got V-1, but I'd like to describe the reanson for that later
08:15:22 <takashi> I tested in my all in one vm, and confirmed it works
08:16:10 <takashi> I need to test again because I rebased the patch sometime after that testing, but I think I already finished about the essential changes.
08:16:44 <takashi> I still have some remaining items, which is described in my comment partially
08:16:59 <takashi> 1. Update docs: We have manuall installation doc, and I need to update it
08:17:17 <takashi> 2. Fix ansible playbooks which is run in 'ant deploy'
08:17:28 <takashi> 3. Fix ansible playbooks for uninstalling
08:17:36 <takashi> 4. Test in-docker installation
08:18:13 <takashi> 5. Fix gate job timeout caused by slow copy of the directory
08:18:53 <takashi> eranrom: do the above items make sense to you?
08:19:07 <eranrom> takashi: absolutely. few comments:
08:19:36 <eranrom> 1. This is probably the largest "big tent" task that you have taken upon yourself to thank you very much!
08:20:03 <eranrom> 2. I will be happy to do the doc update after an "initial landing"
08:20:35 <eranrom> 3. If you are not using the "ant deploy" on an everyday basis (I do) I will be happy to do #2 above as well.
08:20:39 <eranrom> 4. end of comments
08:21:28 <takashi> eranrom: Do you think we had better keep ansible playbooks for 'ant deploy' and uninstalling?
08:21:48 <takashi> currently it is not used at least for our development/testing
08:22:19 <eranrom> uninstalling is definately not used and probably outdated, so we better off just get rid of them
08:22:32 <takashi> eranrom: +1
08:22:56 <eranrom> I use ant deploy.
08:23:23 <takashi> eranrom: ok
08:23:58 <eranrom> I find it convenient to deploy changes during dev. I am curious to know how you are working. Say you change Sbus stuff that needs to go into the container
08:24:03 <eranrom> what do you do?
08:24:34 <eranrom> Hi Kota
08:24:37 <kota_> hello
08:24:46 <takashi> kota_: hi
08:24:47 <kota_> sorry for late
08:24:57 <eranrom> kota_: np good to have you
08:25:07 <kota_> :-)
08:25:31 <takashi> Is it better to back to design summit topic?
08:25:32 <kota_> let me check current meeting status
08:25:59 <takashi> kota_: we finished design summit topic, and talking about big tent updates
08:26:05 <kota_> k
08:26:14 <takashi> kota_: about updates about package work
08:26:21 <kota_> nice
08:26:59 <eranrom> takashi: sure, lets just finish this one. so, if you are not using the deploy, I can also update them. We can discuss alternatives offline
08:27:09 <takashi> eranrom: When I test functional environment, I set up the whole environment using ansible
08:27:32 <eranrom> I see, so most work is done using unittests?
08:27:33 <takashi> I'll have a look about the playbook used in ant deploy later.
08:27:41 <takashi> eranrom: yes
08:27:49 <eranrom> takashi: ok gotcha.
08:27:59 <eranrom> Should be get back to design summit?
08:28:08 <takashi> My concern is that we currently have independent ansible playbooks
08:28:19 <takashi> for installing whole environemnt and for installing dev env
08:28:26 <takashi> I'll check if I can merge them
08:28:46 <takashi> to let us maintain them easily
08:28:52 <eranrom> takashi: ok, as far as I recall, they do share some roles, but perhaps it can be improved.
08:29:27 <takashi> eranrom: Let's go back to duscussion about Design summit. It's better to decide the schedule
08:29:35 <eranrom> takashi: agree.
08:29:35 <takashi> as early as possible
08:29:38 <kota_> ok
08:29:49 <kota_> wow, is it possible?
08:30:57 <eranrom> kota_: well, we are a small team, and even if we are not allocated rooms we can find some spot to talk.
08:31:17 <eranrom> We may get something on Fri, but then it should be for subjects where Kota can be absent
08:31:23 <kota_> exactly
08:31:42 <eranrom> So I guess the main question is, when do we want to meet during the week.
08:32:27 <eranrom> For me, I have a talk on Thu. noon (~13:00), and we may have a fishbowl in the morning. Otherwise, I plan to spend most of the time in Swift's design sessions.
08:32:58 <takashi> kota_: FYI, currently we can have a room for fb session in Thurs. morning, and working room on Fri.
08:33:23 <takashi> as described in the meeting agenda
08:34:01 <kota_> k, sounds like, we can manage a mini meeting in Tue or Wed?
08:34:25 <eranrom> How about adding to the Etherpad the available times each one has?
08:34:40 <kota_> eranrom: good thought
08:34:48 <takashi> eranrom: +1
08:35:13 <kota_> eranrom: right now i don't check summit schdule in detail except Swift desgin sessions.
08:35:20 <kota_> didn't
08:35:55 <kota_> I think I can manage the time except Swift's slots.
08:36:31 <eranrom> Ok, so I will also put on the etherpad the Swift design summit slots
08:36:57 <eranrom> So here is what I suggest:
08:37:09 <eranrom> 1. We place Swift design summit slots on the etherpad
08:37:13 <kota_> eranrom: ok, will confirm other schedule in this week.
08:37:20 <eranrom> 2. Add personal conttaints to the etherpad
08:37:45 <eranrom> 3. populate etherpad with working session subjects
08:38:01 <eranrom> 4. next IRC meeting we can decide which topic to discuss when
08:38:11 <eranrom> end of suggestion :-)
08:38:15 <kota_> :-)
08:38:20 <takashi> eranrom: makes sense
08:38:30 <eranrom> ok, so next topic?
08:39:27 <eranrom> #topic Release planning
08:40:45 <takashi> The reason why I added this one is I'm currently wondering if it is better to have 'Newton' release
08:41:07 <takashi> while we are currently working to join to OpenStack official projects
08:41:15 <eranrom> takashi: you mean and skip "otaca"?
08:41:52 <eranrom> takashi: Sorry, confused the order.
08:42:01 <takashi> eranrom: no. AFAIK the other openstack projects has Newton release in the next week (right?).
08:42:07 <takashi> s/in/until/
08:42:32 <takashi> (maybe swift has more early release
08:42:54 <kota_> by
08:44:06 <eranrom> takashi: generally, I think we should, but then I ask myself, what is it that we are releasing.
08:44:59 <eranrom> takashi: Is it a stable branch? if so should we make sure to take specific Swift version on it?
08:45:23 <eranrom> takashi: and specific Keystone version..
08:46:31 <takashi> generally, when we tests master, it is tested against master version of the other project
08:47:14 <eranrom> takashi: right.
08:47:26 <takashi> and when we tests one release, it makes sense to me to test it on the same release of the other project
08:47:36 <eranrom> agree.
08:47:43 <takashi> I'll check gate job for stable release. I think I can find some information about that.
08:48:16 <eranrom> takashi: so you suggest that the release will be a stable branch where we fix the dependency versions?
08:48:38 <takashi> eranrom: ideally
08:48:46 <eranrom> takashi: sounds good to me.
08:49:32 <eranrom> takashi: so I guess the next question is what do we want to land before that. Any thoughts?
08:50:43 <takashi> I'm currently thinking about python storlet work, especially about the remaining documentatnion
08:51:11 <kota_> cool
08:51:17 <eranrom> +1
08:52:07 <eranrom> I now realize that I have skipped "Remaining docs for python storlets [takashi]". Sorry!
08:52:43 <takashi> eranrom: np. I thought we can discuss about that in this topic, so just skipped
08:52:49 <takashi> We need to add some more tests about that, but I think we can release it at least as beta state, if we can add that doc
08:52:51 <eranrom> ok, so please go ahead
08:53:46 <eranrom> takashi: sounds good to me!
08:53:55 <takashi> the only one remaining item about doc part is the guide for developing storlet applications
08:54:34 <eranrom> takashi: Can I help with that?
08:54:55 <takashi> eranrom: That's the thing I would like to bring here
08:55:05 <eranrom> takashi: I will be happy to do that.
08:55:27 <eranrom> it will also help me get a better grip on python storlets.
08:55:46 <takashi> eranrom: Can I ask your help about that? At least, I need a base to split doc for java and doc for python
08:55:54 <eranrom> I can upload an initial patch by the end of the week
08:55:58 <eranrom> takashi: sure!
08:56:34 <takashi> eranrom: thanks for your help. If you have anything you need to confirm about python storlet, let me know
08:56:41 <eranrom> takashi: sure!
08:56:43 <takashi> eranrom: Or I can follow-up TODOs
08:57:07 <eranrom> np. We are running out of time and there was a small thing I wanted to bring up - is it ok?
08:57:16 <takashi> eranrom: ok
08:57:19 <kota_> sure
08:57:23 <eranrom> #topic: Review our review process - should we change anything at this point?
08:57:42 <eranrom> There have been several cases where I was too eager to merge stuff that you were still reviewing
08:58:02 <eranrom> So the question is, until I improve/align my review skills, should we change the process?
08:58:19 <eranrom> e.g. +1 only after 2X +2?
08:58:26 <eranrom> or something else?
08:58:47 <kota_> sorry, i didn't get the suggestion
08:59:09 <eranrom> should we merge something only after we have two reviewers giving +2?
08:59:37 <kota_> ah, that's the openstack style.
08:59:42 <takashi> eranrom: I think that should be a basic rule.
08:59:54 <eranrom> ok, Lets switch to #openstack-storlets. need to end now
09:00:07 <eranrom> #endmeeting