09:00:33 <masahito> #startmeeting blazar
09:00:34 <openstack> Meeting started Tue Jul  3 09:00:33 2018 UTC and is due to finish in 60 minutes.  The chair is masahito. Information about MeetBot at http://wiki.debian.org/MeetBot.
09:00:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
09:00:37 <openstack> The meeting name has been set to 'blazar'
09:01:04 <masahito> #topic RollCall
09:01:07 <hiro-kobayashi> o/
09:01:15 <priteau> o/
09:01:19 <masahito> Today's agenda is
09:01:20 <tetsuro> o/
09:01:20 <hiro-kobayashi> I'm sorry but I will leave in about 30mins
09:01:27 <masahito> 1. Placement support
09:01:33 <masahito> 2. PTG schedule
09:01:35 <masahito> 3. AOB
09:01:44 <masahito> hiro-kobayashi, priteau, tetsuro: hi
09:01:51 <priteau> Hi everyone
09:01:58 <hiro-kobayashi> hello!
09:02:03 <masahito> hiro-kobayashi: got it. Let's start from your topic.
09:02:07 <hiro-kobayashi> ok
09:02:14 <masahito> anything else?
09:02:53 <masahito> #topic Placement API support
09:02:58 <hiro-kobayashi> I have one topic to discuss about the Placement spec: https://review.openstack.org/#/c/576343/
09:03:30 <masahito> hiro-kobayashi added the item. Let's start from this.
09:03:35 <hiro-kobayashi> It is about the ID of the host.
09:04:01 <hiro-kobayashi> To access the resource provider of the compute host, the UUID of the host is neccesary
09:04:50 <hiro-kobayashi> The host UUID is returned by ver.2.53 and later
09:05:03 <hiro-kobayashi> See https://developer.openstack.org/api-ref/compute/#show-hypervisor-details in detail.
09:05:36 <hiro-kobayashi> So, if blazar uses ~2.52 version of Compute API, the id is integer.
09:06:15 <priteau> When was 2.53 released?
09:06:16 <hiro-kobayashi> So, I added a new column 'resource_provider_uuid' into the computehosts table
09:06:48 <hiro-kobayashi> This is for keeping existing host ids.
09:07:01 <priteau> It's part of the Pike release version 9.1.0
09:07:31 <hiro-kobayashi> Alternative is not to add the 'resource_provider_uuid' and update the existing host's ids to the string UUID in a migration script.
09:07:45 <hiro-kobayashi> priteau: yes.
09:08:07 <tetsuro> priteau: 2.53 is pike maximum
09:08:15 <hiro-kobayashi> priteu: In ChameleonCloud, Blazar uses the integer host id, right?
09:08:25 <tetsuro> https://docs.openstack.org/nova/latest/reference/api-microversion-history.html
09:08:33 <hiro-kobayashi> thanks tetsuro
09:09:05 <priteau> hiro-kobayashi: yes. We are running Ocata's Nova.
09:09:21 <hiro-kobayashi> priteau: OK. thanks!
09:09:33 <tetsuro> Today, I come to think we can use hostname as an identifier and we can get the uuid from the placement using the hostname
09:09:39 <priteau> But I don't want the fact that we run an older release to complicate the code
09:10:11 <priteau> hiro-kobayashi: Actually, we create Blazar hosts based on the hypervisor hostname. The lookup of the Nova ID is automatic.
09:10:35 <hiro-kobayashi> tetsuro: Do you mean the 'resource_provider_uuid' column is not needed?
09:10:53 <tetsuro> I mean that is not needed, yes.
09:11:40 <masahito> hiro-kobayashi: My question is latest Hypervisor list API supports query parameter, right?
09:11:58 <hiro-kobayashi> tetsuro: make sense. Use the hostname to access the resource provider.
09:12:38 <tetsuro> Looks like hostname is identical, so we can use that hostname to get the uuid asking to placement every time
09:12:51 <hiro-kobayashi> masahito: yes, it seems to.
09:13:06 <priteau> Looking up from hostname is the most admin-friendly way to do it.
09:14:05 <hiro-kobayashi> OK. Now it's clear. Remove the 'resource_provider_uuid' and use the hostname instead. Then, the version problem is solved.
09:14:30 <hiro-kobayashi> Agree?
09:14:44 <tetsuro> Agreed.
09:14:49 <masahito> agreed.
09:14:51 <priteau> Agreed.
09:15:01 <hiro-kobayashi> Thanks. I'll update the spec.
09:15:17 <hiro-kobayashi> 1 more thing to share about the Placement spec.
09:15:40 <hiro-kobayashi> I changed the primary-assignee to tetsuro
09:15:52 <tetsuro> that's good with me.
09:15:53 <hiro-kobayashi> tetsuro has already started implementation and I think it's better.
09:16:09 <masahito> np
09:16:19 <hiro-kobayashi> Thanks tetsuro! And I added you in co-authored-by
09:16:42 <tetsuro> Thanks!
09:17:18 <priteau> Thanks tetsuro and hiro-kobayashi!
09:17:30 <masahito> Thanks a lot!
09:17:36 <hiro-kobayashi> I'll update the Launchpad assignee, too.
09:17:42 <hiro-kobayashi> My pleasure ;-)
09:17:52 <masahito> I already see lots of patch in gerrit. Great works!
09:18:31 <hiro-kobayashi> That's all about Placement API support
09:19:20 <masahito> hiro-kobayashi: Thanks.
09:19:29 <masahito> anything else?
09:20:03 <hiro-kobayashi> no
09:20:11 <masahito> #topic PTG schedule
09:20:48 <masahito> Thierry shared the proposed schedule of the Denver PTG.
09:20:50 <masahito> Thierry
09:20:58 <masahito> http://lists.openstack.org/pipermail/openstack-dev/2018-June/131881.html
09:21:19 <masahito> Blazar is scheduled on Monday and Tuesday.
09:21:44 <masahito> Does someone have any conflict or problem at the schedule?
09:22:26 <tetsuro> nope, I can join
09:22:32 <masahito> Scientific SIG is also scheduled on Monday.  It is one of the schedule conflict.
09:22:36 <priteau> There's the Scientific SIG at the same time, but I don't know yet if I will attend anyway.
09:23:45 <masahito> OK. looks no big schedule conflict.
09:24:39 <priteau> Having Nova separate is the most important, so that we can attend placement API discussions
09:24:39 <masahito> any comments for the schedule?
09:25:31 <masahito> Yes. I guess that's why the Blazar is scheduled on Mon. and Tue.
09:25:44 <tetsuro> Cool!
09:26:55 <masahito> #topic AOB
09:27:20 <masahito> Does someone have topic to share or discuss?
09:28:00 <priteau> I do
09:28:18 <masahito> priteau: okay, you have a floor :-)
09:29:42 <priteau> I am starting a new job at StackHPC today. I will still be involved in Chameleon for some time.
09:30:02 <priteau> And in Blazar of course :-)
09:30:13 <hiro-kobayashi> Oh, Congrats!
09:30:23 <masahito> Congrats!!
09:31:00 <priteau> Thanks! I have a new email address that I will share with you, though my uchicago.edu address is still working until mid-August.
09:31:17 <hiro-kobayashi> What is your primary role or mission?
09:32:23 <masahito> crossing my fingers to your new job
09:32:38 <tetsuro> Congrats!
09:33:49 <priteau> hiro-kobayashi: My role is Senior Technical Lead. StackHPC offers consulting so I may be working for various clients.
09:34:06 <priteau> Hopefully Blazar will be an important part of it :-)
09:34:46 <hiro-kobayashi> priteau: Nice role! I hope so :-)
09:36:30 <masahito> I hope so. We're waiting nice feedbacks from user side and developer side ;-)
09:37:34 <masahito> I have one item to hear your idea.
09:37:42 <hiro-kobayashi> I'm sorry but I gotta go. thank you and good bye all
09:37:58 <masahito> hiro-kobayashi: thanks
09:38:12 <masahito> TC team asked me whether this team or projct has any problem.
09:38:46 <masahito> They told they would support the issue.
09:39:55 <masahito> IMO, complex document links is one of our problem.
09:40:21 <masahito> any other ideas to ask them?
09:40:59 <priteau> What is the problem with document links?
09:41:03 <masahito> s/ to ask them?/ to ask them for support?/
09:42:16 <masahito> I means docs.openstac.org or openstack site have lots of link to specific project but the configuration is very complex.
09:42:46 <masahito> So it's hard to update all the link correctly.
09:44:30 <masahito> I plan to tell which repoes we should update?
09:45:06 <masahito> For example, the project navigator doesn't show Blazar project even though we're an official but I can't find the repo for the page.
09:45:31 <priteau> I see. I think some things on the web page may be managed outside of Git repos.
09:46:02 <masahito> right.
09:47:41 <masahito> anything else? if nothing, we can finish early this meeting.
09:48:11 <tetsuro> nothing from me.
09:49:20 <priteau> Nothing else from me.
09:49:21 <masahito> OK, let's finish the meeting.
09:49:21 <masahito> thanks all!  Next weekly meeting has review meeting, too.
09:49:31 <tetsuro> Thanks
09:49:35 <masahito> bye.
09:49:35 <masahito> #endmeeting