08:05:45 <kota_> #startmeeting openstack-storlets
08:05:46 <openstack> Meeting started Tue Nov 22 08:05:45 2016 UTC and is due to finish in 60 minutes.  The chair is kota_. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:05:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
08:05:49 <openstack> The meeting name has been set to 'openstack_storlets'
08:06:09 <kota_> k, let get started
08:06:23 <takashi_> It seems that eranrom already created agenda for today's meeting https://wiki.openstack.org/wiki/Meetings/Storlets
08:06:37 <kota_> #link  https://wiki.openstack.org/wiki/Meetings/Storlets
08:07:04 <kota_> the first item looks about PTG
08:07:23 <takashi_> kota_: yes
08:07:28 <kota_> i think eranrom has been working to propose the slots to discuss in the PTG
08:07:55 <kota_> and unfortunately, it cannot make in Mon-Tue but Wed-Fri.
08:08:06 <kota_> so completely over the Swift PTG.
08:08:10 <takashi_> it seems that we need to be allocated from Wed-Fri, because we are currently proposing storlets vertical (not cross-project horizonal) session
08:08:29 <kota_> yeah, not cross-project
08:08:37 <takashi_> is there any schedule already available now?
08:09:17 <kota_> I'm feeling it's a bug of ptg, if we consider to work a couple of projects, the schedule constraint completely make us bad to work for both.
08:09:44 <takashi_> kota_: yes
08:09:46 <kota_> takashi_: i don't think so
08:09:54 <kota_> just...
08:10:06 <takashi_> it seems we can not find any more detailed information... :-( http://www.openstack.org/ptg
08:10:22 <kota_> we can see the project names which are available to attend in ptg
08:11:01 <kota_> anyway, eranrom is asking a couple of ways we can make the storlets slots
08:11:31 <kota_> A. in parallel with swift or B making a place for storlets in the Swift ptg
08:12:02 <kota_> how do you think of?
08:12:17 <takashi_> kota_: IMO, to avoid schedule collision against swift schedule, which may be a big problem for us,B makes sense
08:12:44 <takashi_> eranrom: hi
08:12:45 <kota_> hi, eranrom
08:12:48 <openstack> eranrom: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
08:12:53 <akihito> hi
08:13:00 <takashi_> eranrom: it is already started!
08:13:03 <eranrom> Hi, Sorry I am late
08:13:08 <kota_> no worries
08:13:14 <takashi_> eranrom: np :-)
08:13:24 <eranrom> Thanks for starting
08:13:27 <kota_> oh, do I have to set the meeting name as storlets? I did openstack-storlets
08:13:45 <eranrom> That's fine.
08:13:49 <kota_> kk
08:14:05 <takashi_> eranrom: we are following your agenda, and just talking about ptg stuffs
08:14:18 <eranrom> cool.
08:15:04 <eranrom> Looking for the logs...
08:15:34 <kota_> takashi: i'm with you, at the this time, we could manage our spaces into swift ptg rather than locate different spaces
08:15:58 <kota_> takashi: because it could make difficult to work together with swifters
08:18:33 <takashi_> kota_: yes. I'd like to ask eranrom's opinion, but it seems better to me to ask notmyname to allow us to allocate storlets session in swift session timeframe.
08:18:39 <takashi_> to work with swift community
08:18:39 <kota_> and i think, Swift ptg will work as well as past design summit (or mid-cycle) so we can manage our time to discuss storlets in the swift slots
08:19:08 <takashi_> kota_: just thinking the same thing
08:19:46 <kota_> or I'm wondering if we could place concatenated room for such collaborate prjects :/
08:19:58 <kota_> sure
08:20:17 <eranrom> Originally, I also thought that we should work with Swift this time, however, acoles told me that given the shorter period of PTG compared to the previous mid cycle
08:20:25 <eranrom> 3 days instead of 5
08:20:30 <eranrom> Its going to be very busy
08:20:59 <eranrom> I can ask John though
08:21:08 <kota_> eranrom: exactly
08:21:54 <kota_> or if we could make *un-official* slots in Mon-Tue?
08:22:17 <kota_> tc may not want to do that though :/
08:22:17 <eranrom> You mean just be there, and look for a room.
08:22:24 <eranrom> Sounds lije a good idea to me.
08:22:25 <kota_> eranrom: sure
08:22:26 <eranrom> like
08:23:07 <kota_> for this time, I'm going to be in the whole week there.
08:23:21 <takashi_> kota_: same for me
08:23:26 <eranrom> I guess we need to choose between two non optimal choices:
08:23:30 <kota_> and probably not so busy in Mon-Tue.
08:23:39 <eranrom> 1. Do this Mon/Tue unofficially.
08:23:59 <eranrom> 2. Do this Wed-Fri where both you and Takashi are very busy with other projects
08:24:16 <eranrom> Seems to me that option 1 is better. What do you think?
08:24:46 <eranrom> We may not be allocated with a room, but I believe we can manage
08:25:19 <kota_> 1 seems ok. And try 2 if we have to consider about swift upstream.
08:26:00 <kota_> and 3. make a feedback ''
08:26:31 <eranrom> kota_: In 2, Do you mean if we have storlet related issues that are related to Swift?
08:26:32 <kota_> to enable to attend more than one projects
08:26:42 <kota_> eranrom: yes
08:27:12 <takashi_> talking about my situation, I have some topics in nova and cinder, which I should join discussion about.
08:27:29 <kota_> in Wed-Fri?
08:27:34 <kota_> takashi_:^^
08:27:39 <takashi_> kota_: maybe
08:27:47 <eranrom> takashi_: Does this mean that you are also in favor of option 1?
08:27:55 <kota_> ok, so 1 seems to work well at this time.
08:28:18 <takashi_> so if we have session in Wed-Fri, it is difficult to fix the timeframe when we have room, because there are no available schedule about nova and cinder
08:29:07 <takashi_> so if we have Wed-Fri, it works better for me to let us allocate in swift session, which is expected to have relatively long timeframe,
08:29:48 <takashi_> because we can choose the meeting schedule ad-hocly, based on the chaning situation.
08:30:07 <takashi_> but I don't have any big topics in Mon-Tue, so still 1 works fine for me
08:30:25 <takashi_> s/chaning/changing/
08:30:59 <eranrom> ok, so lets make it Mon-Tue unofficial (perhaps we need only one day) and then come with feedback
08:31:42 <eranrom> There is one more thing I forgot to mention (which may not be relevant given that we chose the unofficial 'path') and that is
08:31:53 <eranrom> Weneed to be official so that we can get allocated room
08:32:22 <eranrom> I was told that rooms are being kept for us (Wed-Fri) under the assumption that we will be official by Feb.
08:32:33 <eranrom> I do not see a reason why that would not happen
08:33:31 <kota_> yea
08:34:00 <eranrom> anything else on this topic?
08:35:05 <takashi_> eranrom: just a little. maybe there are not so many things already decided about ptg scheduling.
08:36:01 <takashi_> eranrom: so another possible option is that we basically ask rooms for ptg, and then discuss about the desired schedule, which works well for us, with tc, ptls(john) or other foundation members
08:37:46 <eranrom> takashi_: I have asked ttx if we could have rooms Mon/Tue. He replied that he is keeping rooms for us (assuming we get official by Feb.), but these would be only Wed-Fri
08:38:46 <takashi_> eranrom: do we have a room for all Wed-Fri? or have a room for limited timeframe in Wed-Fri.
08:38:56 <takashi_> ?
08:39:15 <eranrom> takashi_: My understanding is that its for all Wed-Fri
08:39:35 <eranrom> I can validate this.
08:40:00 <eranrom> Does this make you think that option 2 is better?
08:41:24 <takashi_> eranrom: will think a little more later. Can I ask you to available time long, currently expected?
08:41:45 <takashi_> because we have another big items around installation/packaging work...
08:42:11 <eranrom> ok, we can continue this afterwards in #openstack-storlets
08:42:13 <takashi_> I mean, I'd like to confirm that they are really proposing whole Wed-Fri
08:42:19 <takashi_> eranrom: yes
08:42:21 <eranrom> takashi_: will do.
08:42:41 <takashi_> eranrom: thx
08:42:42 <eranrom> topic: installation/packaging work
08:42:47 <eranrom> #topic: installation/packaging work
08:42:58 <eranrom> #topic installation/packaging work
08:43:00 <takashi_> I suppose that kota_ should do that
08:43:05 <eranrom> :-) right
08:43:09 <kota_> #topic: installation/packaging work
08:43:10 <takashi_> because he started this meeting
08:43:14 <eranrom> Thanks! :-)
08:43:18 <kota_> sure
08:43:31 <kota_> i didn't know the rule though
08:44:08 <eranrom> kota_: no worries, it just makes it easier to browse the logs afterwards
08:44:26 <kota_> #link https://review.openstack.org/#/c/370332/
08:44:37 <eranrom> takashi_: Do you want to take over?
08:45:01 <kota_> takashi_: ?
08:45:12 <takashi_> so currently I finally make package patch pass gate job, making it dependent on eranrom's devstack patch
08:45:29 <kota_> yeah, that's awesome
08:45:45 <eranrom> +1!
08:46:24 <takashi_> I'd like to discuss about two things today
08:46:37 <takashi_> 1) Is there anything remaining about devstack patch?
08:46:56 <takashi_> 2) What should be done before merging packaging patch
08:47:23 <takashi_> for 1, basically I'd like to ask kota_'s opinion
08:47:35 <kota_> i think 1 is for eranrome?
08:47:52 <kota_> i'm seeing eranrom has updated the devstack patch.
08:48:02 <eranrom> Right. For 1, I have uploaded a new patch yesterday, addressing some of Kota's remarks.
08:48:38 <eranrom> Its ok from Jenkins P.O.V. but I still need to tests the 'dev' flavour.
08:48:54 <eranrom> After validating this, I would be happy to get that merged.
08:49:15 <kota_> k, and in parallel, I will review again the changes.
08:49:15 <takashi_> eranrom: so you mean that it is better to wait for your testing, right?
08:49:24 <kota_> maybe in Thu or Fri in this week.
08:49:34 <eranrom> right, I will do so today, and ping you once done
08:49:45 <kota_> thx eranrom
08:49:54 <takashi_> eranrom: thanks. I'll review again, and test by myself, if possible
08:50:29 <eranrom> takashi_: kota_: great, thanks. Let me ping you before you review / test as I might need an additional patch for that,
08:50:48 <eranrom> that is an updated patch
08:50:58 <takashi_> eranrom: After you fix you patch, I'll rebase my patch on your update. I think it does not require so much effort
08:51:05 <kota_> for item 2, i think my concerns can be almostly in follow-up
08:51:21 <takashi_> kota_: ok
08:51:38 <kota_> but still only item i'd like to make further work is 'describing how we can install/deploy in the new package'
08:52:10 <kota_> takashi_: you had written already some notes in commit message
08:52:29 <kota_> but still not describe concretely that yet, right?
08:52:39 <takashi_> kota_: you are right
08:52:46 <kota_> I'm recently getting how it works via reviewing
08:52:59 <kota_> maybe, in the new package, we need a couple of commands
08:53:07 <kota_> do ./install_lib.sh
08:53:16 <kota_> and do python setup.py install
08:53:23 <takashi_> kota_: yes
08:53:47 <kota_> that seems awesomely progressed rather than ansible script. How easy to deply!!!
08:53:57 <kota_> but I'd like to add the docs for anyone.
08:54:24 <kota_> hopefully, I'd like to make it into one script but that can be in follow up
08:54:44 <kota_> takashi_: does it make you sense?
08:55:00 <takashi_> kota_: yes, but some thoughts from my side
08:55:32 <takashi_> talking about doc, currently we have manual installation guide, which follows all procedure executed by ansible.
08:56:09 <takashi_> and I don't know whether we should keep current thing, considering current discussion for splitting ansible script
08:57:00 <takashi_> The other thing is that for python package, which I think storlets should be designed as, I think we do not wrap 'python setup.py'.
08:57:31 <takashi_> but all things can be discussed in follow-up, if you are ok about them. just recording my thought
08:58:03 <takashi_> s/do/should/g
08:58:04 <eranrom> I think we need to wrap up. Do you want to move to #openstack-storlets?
08:58:32 <takashi_> eranrom: yes, though I don't have so long time
08:58:33 <kota_> for the first item, we could say 1. installation 2.deploy whole parts including swift as different clearly. IMO
08:58:46 <kota_> yeah, we should move
08:58:58 <kota_> #endmeeting openstack-storlests