04:00:48 <tpatil> #startmeeting Masakari
04:00:49 <openstack> Meeting started Tue Dec  5 04:00:48 2017 UTC and is due to finish in 60 minutes.  The chair is tpatil. Information about MeetBot at http://wiki.debian.org/MeetBot.
04:00:50 <rkmrHonjo> tpatil: thanks.
04:00:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
04:00:53 <openstack> The meeting name has been set to 'masakari'
04:01:01 <tpatil> #link https://wiki.openstack.org/wiki/Meetings/Masakari
04:01:19 <tpatil> #topic High priority items
04:02:10 <tpatil> Any one wants to bring up any high priority items?
04:02:29 <sagara> no
04:02:50 <tpatil> Moving to next topic
04:03:00 <tpatil> #topic Bugs (stuck/critical)
04:03:19 <rkmrHonjo> #link https://review.openstack.org/#/c/486576/
04:03:40 <rkmrHonjo> The above patch addresses py35 UT error. Please review it.
04:04:22 <tpatil> rkmrHonjo: I will review it
04:04:28 <rkmrHonjo> tpatil: thanks.
04:05:05 <tpatil> I have seen Takahara has proposed a new PS : https://review.openstack.org/#/c/515013/
04:05:11 <tpatil> I will review above PS as well
04:05:43 <tpatil> No new bugs reported
04:06:16 <tpatil> Moving ahead
04:06:20 <tpatil> #topic Discussion points
04:06:29 <tpatil> Horizon Plugin
04:06:47 <tpatil> I can see masakari-dashboard repo is created
04:06:51 <tpatil> #link https://github.com/openstack/masakari-dashboard
04:07:17 <tpatil> Niraj will propose cookie-cutter patch soon
04:08:02 <tpatil> Install guide document (in progress)
04:09:13 <tpatil> Need another +2 and workflow to get all patches merge
04:09:29 <tpatil> Please review these patches
04:10:52 <rkmrHonjo> tpatil: I'll review it if I can. But I'm very busy now, sorry...
04:11:10 <tpatil> rkmrHonjo: Ok
04:11:23 <tpatil> Recovery method customization
04:11:36 <tpatil> #link https://review.openstack.org/#/c/458023/
04:12:10 <tpatil> Please review the specs
04:12:37 <tpatil> There were few review comments from Abhishek which has been fixed
04:13:22 <tpatil> Find hosts without specifying segments
04:13:48 <tpatil> #link https://review.openstack.org/#/c/506986/
04:13:59 <tpatil> I will review the updated specs in this week
04:14:20 <rkmrHonjo> tpatil: thanks.
04:15:01 <tpatil> Queen work items
04:15:23 <tpatil> No updates from my end
04:15:53 <rkmrHonjo> me too.
04:16:09 <tpatil> Ansible support for Masakari
04:16:23 <tpatil> Last week, we discussed about it
04:16:47 <tpatil> Dinesh will start understanding how to develop ansible plugins
04:16:50 <Dinesh_Bhor> No updates for ansible support from my side yet. I have strated exploring ansible plugin for magnum.
04:17:24 <Dinesh_Bhor> I think we have to follow the same for masakari
04:18:34 <rkmrHonjo> Dinesh_Bhor: It sounds good to me.
04:18:47 <tpatil> Dinesh_Bhor: Thanks
04:19:07 <tpatil> Last time, Sampath asked us to create a new blueprint in Ansible project
04:19:46 <tpatil> once we have basic understanding about Ansible plugins, let's submit a new BP in Ansible project
04:19:58 <Dinesh_Bhor> tpatil, yes
04:21:57 <tpatil> Ok, that's end of the agenda
04:22:55 <rkmrHonjo> tpatil: I added a item to Discussion points.
04:22:57 <tpatil> Anyone wants to bring up any new topic?
04:23:05 <tpatil> Let me check
04:23:12 <rkmrHonjo> That is "Call Force down API when host-failure will be notified".
04:23:40 <rkmrHonjo> Can I explain about this item?
04:23:55 <tpatil> Yes ,please
04:24:13 <rkmrHonjo> ok. Current masakari will wait for 3 minutes when host-failure will be notified. ("3 minutes is configurable.)
04:24:28 <rkmrHonjo> The purpose of this is to wait nova-compute down.
04:24:32 <tpatil> Does masakari-monitor wait for 3 mins before notifying to masakari-api?
04:25:06 <rkmrHonjo> tpatil: no. masakari-monitor doesn't wait. masakari wait.
04:25:13 <rkmrHonjo> So, I think, there is a possibility that using force down API instead of waiting make recovery process faster.
04:26:22 <rkmrHonjo> Updated forced down API of nova can change service state of nova-compute.
04:27:25 <Dinesh_Bhor> rkmrHonjo: you mean UP/DOWN ?
04:27:36 <rkmrHonjo> Dinesh_Bhor: no, only DOWN.
04:27:52 <Dinesh_Bhor> rkmrHonjo: yeah
04:28:05 <tpatil> rkmrHonjo: Can you point out the place where masakari waits for 3 mins?
04:28:40 <tpatil> You mean config option: wait_period_after_service_update
04:28:42 <rkmrHonjo> tpatil: ok, please wait a thile.
04:28:47 <rkmrHonjo> s/thile/while
04:29:08 <rkmrHonjo> #link https://github.com/openstack/masakari/blob/93556d3116eb8c62bf881da0edd05692c9ebd66e/masakari/engine/drivers/taskflow/host_failure.py#L52
04:30:20 <tpatil> Ok, that's a good idea. But does it guarantee that nova-compute won't change instance state after force down api returns
04:30:55 <tpatil> I haven't checked force down api yet. will check it at my end
04:31:17 <rkmrHonjo> tpatil: thank you for pointing out. I check it after that, too.
04:31:27 <Dinesh_Bhor> rkmrHonjo: so this new API will *DOWN* the state as well as *Disable* the status of nova-compute service, right?
04:31:58 <Dinesh_Bhor> rkmrHonjo: okay, I wiill check the API by myself. thanks for suggestion.
04:32:24 <tpatil> basically you cannot evacuate any instances until nova-compute service is down on the failed compute host
04:32:33 <rkmrHonjo> Dinesh_Bhor: Call disable API will be remained. I will add calling update forced down API instead of eventlet.sleep.
04:32:57 <rkmrHonjo> taptil: right.
04:32:59 <tpatil> so you can call this api to disable nova-compute service immediately
04:33:17 <Dinesh_Bhor> rkmrHonjo: okay, thanks
04:33:47 <tpatil> rkmrHonjo: Please propose lite specs for this change
04:33:55 <rkmrHonjo> tpatil: ok.
04:34:06 <tpatil> rkmrHonjo: Thank you
04:35:09 <rkmrHonjo> That is all from my side about this item.
04:36:19 <tpatil> Let's end this meeting
04:36:25 <tpatil> Thank you, All
04:36:31 <Dinesh_Bhor> thank you all
04:36:37 <rkmrHonjo> thanks.
04:36:39 <sagara> Thank you
04:36:42 <tpatil> #endmeeting