13:08:44 <Qiming> #startmeeting senlin
13:08:45 <openstack> Meeting started Tue Jun 27 13:08:44 2017 UTC and is due to finish in 60 minutes.  The chair is Qiming. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:08:46 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:08:48 <openstack> The meeting name has been set to 'senlin'
13:09:00 <Qiming> we can get this conversation recorded, :)
13:09:09 <xinhuili> :)
13:09:13 <Qiming> Senlin is part of VIM now?
13:09:20 <xinhuili> not yet
13:09:28 <Qiming> I'm happy to know any feedback
13:09:32 <xinhuili> I am exploring the chance
13:09:44 <Qiming> where we need to improve, where we need to change
13:09:50 <xinhuili> to expose recovery action to the higher level
13:09:54 <xinhuili> control flow
13:10:03 <Qiming> that makes sense
13:10:10 <xinhuili> I see
13:10:20 <xinhuili> you always open to any
13:10:20 <Qiming> seems to me the recover action of workflow is not fixed in senlin yet
13:10:33 <xinhuili> oh
13:10:35 <Qiming> should be request driven now
13:10:52 <xinhuili> I will check the bug
13:11:01 <xinhuili> In ONAP
13:11:24 <Qiming> there are always a thousand new features to develop
13:11:29 <xinhuili> there are projects named as policy and DCAE data collection and analysis engine
13:11:50 <xinhuili> they handle the close loop
13:12:13 <Qiming> DCAE will collect VNF failure events?
13:12:19 <xinhuili> yes
13:12:47 <Qiming> is it agent based or agentless?
13:12:54 <xinhuili> VES
13:12:57 <xinhuili> based
13:13:11 <xinhuili> VES is an OPENFV project
13:13:19 <xinhuili> agent based
13:13:20 <Qiming> sigh ...
13:13:31 <xinhuili> reinvent the wheel?
13:13:54 <Qiming> I'm never a big fan of a huge collection of software
13:14:09 <xinhuili> I see
13:14:39 <xinhuili> I am thinking to adopt senlin for the cluster.do functions
13:14:51 <Qiming> okay
13:14:53 <xinhuili> but the loop is so big
13:15:03 <xinhuili> I do not know if we can implement it in this release
13:15:36 <Qiming> one thing I recently read is about solution space and problem space
13:15:49 <xinhuili> tell me more :)
13:16:18 <Qiming> there are always people trying to solve a problem in a solution space
13:16:40 <Qiming> few people are trying to find a solution from a problem space
13:17:19 <Qiming> can you feel the difference?
13:18:47 <xinhuili> hammer nail
13:19:02 <Qiming> the former group of people favor integrating a lot of software components to achieve a goal
13:19:25 <Qiming> the latter group of people focus more on what is good enough to solve the problem
13:19:36 <Qiming> yes
13:20:44 <xinhuili> :)
13:21:15 <Qiming> another survey I read is about why people are pursuing a phd degree
13:21:24 <xinhuili> :)
13:21:58 <xinhuili> why
13:22:05 <xinhuili> top three reasons?
13:22:06 <Qiming> I was shocked by the result: only 5% of the candidates are really interested in academic research
13:22:20 <xinhuili> at least you are
13:22:29 <xinhuili> so you are shocked
13:22:33 <xinhuili> joking
13:22:54 <Qiming> I'm trying to comparing that to a open-source community, :)
13:23:03 <xinhuili> what do the people want
13:23:16 <xinhuili> haha
13:24:17 <Qiming> how many people are really interested in building a good opensource software
13:24:47 <Qiming> you can do an analysis in the ONAP community as well, :)
13:25:05 <xinhuili> :)
13:25:41 <xinhuili> I saw xuefeng
13:25:50 <xinhuili> XueFengLiu
13:26:09 <XueFengLiu> hi
13:26:14 <xinhuili> hi
13:26:17 <Qiming> hi, XueFengLiu
13:26:32 <Qiming> thanks for joining
13:26:45 <Qiming> any news from your side?
13:27:33 <XueFengLiu> yes
13:28:13 <XueFengLiu> rdo is in process by a new colleague
13:28:20 <Qiming> cool
13:28:37 <xinhuili> rdo
13:28:41 <xinhuili>13:28:51 <xinhuili> sorry for my long lose
13:28:55 <rbowen> RPM Distribution of OpenStack - http://rdoproject.org/
13:28:58 <XueFengLiu> And two person research  TIGK
13:29:14 <XueFengLiu> for monitoring
13:29:17 <Qiming> getting senlin released under RDO
13:29:51 <xinhuili> cool
13:30:07 <XueFengLiu> yes, we need rpm for senlin
13:30:19 <Qiming> TIGK?
13:30:58 <XueFengLiu> a framework for monitoring
13:32:16 <Qiming> never heard of that, an opensource solution?
13:33:06 <XueFengLiu> yes
13:33:40 <XueFengLiu> http://git.openstack.org/cgit/openstack/openstack-ansible-ops/
13:33:51 <XueFengLiu> here is the install for them
13:34:37 <XueFengLiu> and for senlin and zaqar
13:36:19 <Qiming> alright, G-Grafana, I-Influx, K-Kapacitor ... what is T?
13:36:40 <Qiming> Telegraf
13:36:43 <Qiming> got it
13:36:52 <XueFengLiu> yes
13:37:11 <XueFengLiu> now ,we only research and comparison
13:37:17 <Qiming> okay
13:37:24 <Qiming> I was chatting with xinhui about this
13:37:29 <XueFengLiu> and xinhui's video, I have watched
13:37:38 <Qiming> If I can make a suggestion
13:38:02 <Qiming> I would start from the requirements at hand, instead of from the software available
13:38:52 <Qiming> I have recently shared a paste on wechat, about "not everyone should be a google"
13:39:12 <Qiming> hope that article can be of help
13:40:04 <XueFengLiu> so it means?
13:41:27 <Qiming> it is an article about the enthusiasm about map-reduce and hadoop during the past 10 years
13:41:59 <Qiming> more and more people found that hadoop is not suitable for their usage scenario
13:42:33 <Qiming> it is not hadoop's fault, it is the user's fault
13:42:50 <ruijiej> that costs too much:)
13:43:19 <Qiming> thousands of papers got published, anyway
13:44:14 <Qiming> you will come up with a good solution if your focus is on the problem you want to solve
13:44:35 <XueFengLiu> yes
13:44:46 <XueFengLiu> that's right:)
13:44:53 <Qiming> however, if you focus too much on how famous, how hot a software is today, ...
13:45:08 <XueFengLiu> en
13:45:32 <Qiming> hope I'm not worrying too much, :)
13:46:32 <Qiming> ruijiej, any update from your side?
13:46:35 <XueFengLiu> yes, got it
13:47:55 <Qiming> alright, ruijiej is snoring ...
13:48:29 <ruijiej> :)
13:48:40 <XueFengLiu> Qiming, research these is not by me. it's a team:) And I am continue thing about how to use senlin in the framework.
13:48:50 <ruijiej> sorry, I am on a trip:)
13:49:37 <XueFengLiu> thinking
13:49:41 <Qiming> okay, feel free to propose any improvements
13:50:14 <XueFengLiu> ok
13:52:02 <Qiming> anything else you want to talk about before we release the channel?
13:52:22 <XueFengLiu> for network can senlin do more things?
13:52:33 <Qiming> yes
13:53:04 <Qiming> e.g. building pools for network resources
13:53:21 <Qiming> e.g. simplifying the network configuration of a cluster
13:53:43 <XueFengLiu> then we can control this resources
13:54:29 <XueFengLiu> so we can have more discuss for this
13:54:38 <Qiming> the goal is not to control them, the goal, in my opinion, is to make users life easier
13:54:42 <Qiming> sure
13:54:59 <XueFengLiu> ok
13:55:09 <ruijiej> inregrade
13:55:36 <Qiming> better, ease of use abstraction for users
13:55:44 <XueFengLiu> yes
13:55:45 <ruijiej> With other platform or component is also interested?
13:55:56 <Qiming> so they can focus on only things they need to change
13:56:20 <Qiming> yup, ruijiej
13:56:33 <XueFengLiu> so we can abstraction and intergate with skydive
13:56:45 <Qiming> if you can brige senlin to libcloud, for example, :)
13:56:46 <XueFengLiu> abstract
13:57:22 <ruijiej> Okay Qiming, cause we are using something to control k8s"s compute node, hope Senlin can do that directly
13:57:40 <Qiming> I see
13:57:49 <Qiming> shouldn't be difficult
13:58:02 <Qiming> I'm learning myself k8s recently
13:58:23 <Qiming> k8s has python client as well
13:58:31 <XueFengLiu> ok
13:58:51 <XueFengLiu> Qiming, we need more discuss
13:58:53 <ruijiej> Thats interesting... I am looking at it's network arch.
13:59:00 <Qiming> then cluster autoscaler won't be a thing you can only do on google compute cloud
13:59:01 <XueFengLiu> ruijiej
13:59:07 <Qiming> you can do that on openstack, :)
13:59:14 <ruijiej> Yes XueFengLiu
13:59:20 <Qiming> time's up, buddy
13:59:26 <Qiming> switch to #senlin pls
13:59:31 <Qiming> #endmeeting