Monday, 2016-10-10

*** thorst has joined #openstack-watcher01:13
*** thorst has quit IRC01:22
*** licanwei has joined #openstack-watcher01:26
*** thorst has joined #openstack-watcher01:31
*** thorst has quit IRC01:32
*** thorst has joined #openstack-watcher02:27
*** thorst has quit IRC02:27
*** thorst has joined #openstack-watcher03:36
*** thorst has quit IRC03:41
*** thorst has joined #openstack-watcher04:38
*** thorst has quit IRC04:46
*** kuntumtarm has joined #openstack-watcher06:07
*** diga has joined #openstack-watcher06:32
*** kuntumtarm has quit IRC06:34
*** thorst has joined #openstack-watcher06:48
*** thorst has quit IRC06:56
*** alexchadin has joined #openstack-watcher07:05
*** dtardivel has joined #openstack-watcher07:14
*** jed56 has joined #openstack-watcher07:35
*** alexchadin has quit IRC07:37
*** alexchadin has joined #openstack-watcher07:37
*** vincentfrancoise has joined #openstack-watcher07:39
*** thorst has joined #openstack-watcher07:54
*** Kevin_Zheng has joined #openstack-watcher07:59
*** thorst has quit IRC08:02
*** thorst has joined #openstack-watcher08:59
openstackgerritVincent Françoise proposed openstack/watcher-specs: Specs for watcher-notifications-ovo
*** thorst has quit IRC09:06
*** Guest10079 is now known as zigo09:23
*** alexchadin has quit IRC09:26
*** alexchadin has joined #openstack-watcher09:26
digaacabot: Hi09:28
digaalexchadin: vincentfrancoise : Hi09:28
digaall given comments are resolved & submitted updated PS on friday09:29
*** alexchadin has quit IRC09:31
*** alexchadin has joined #openstack-watcher10:39
*** thorst has joined #openstack-watcher11:08
*** thorst has quit IRC11:16
alexchadinhello jed5611:39
jed56Hello alexchadin11:39
jed56как вы?11:39
alexchadinЯ отлично. А вы?:)11:39
alexchadinjed56: are you learning Russian language?:)11:40
jed56no but a good friend of me is russian11:41
alexchadinjed56: oh, that's great! It is pretty hard to learn language even for native.11:42
jed56okay ;-)11:42
alexchadinjed56: do you have about 10 minutes to discuss improvements of action plan storage? I'm working on the spec for now.11:43
jed56yes of course11:43
jed56did you push a review ?11:44
alexchadinreview on what?11:44
jed56on the action-plan-storage blueprint ?11:44
alexchadinNot yet:)11:44
alexchadinI've discussed with our architecture about action weights, he said that all the migrations can be launched at the same time for specified action plan.11:47
alexchadinso, we can leave next_action field11:47
alexchadinif we want to specify special subgraph11:48
jed56alexchadin: IMHO, this is not a good idea because if you have 20 migrations at the same time you can saturate the network11:48
alexchadinWe will limit them11:48
jed56Action weight is a simple implementation of the planner, it has many drawback11:49
alexchadinEvery action may have priority11:51
*** thorst has joined #openstack-watcher11:51
alexchadinwe can run actions in parallel in order of these priorities11:51
jed56I  have a meeting can we continue in 30 minutes ?11:53
sballe_jed56: did you take a look at the slides for the summit? Please send me your edits soon11:53
jed56Yes a little bit :)11:53
alexchadinjed56: yeah, of course11:53
acabotdiga : your last PS fail because you have removed references11:54
acabotdiga : please run tox -edocs before submitting your PS for review11:55
digaacabot: you said "remove the references" so I did, but you want me to revert back11:56
digafor b-com11:56
acabotdiga : I didn't say remove references11:57
digaacabot: I means, b-com link11:57
acabotdiga : I said "all links to are dead links, please use the official doc now on"11:57
acabotdiga : keep the links but replace by docs.openstack.org11:57
digab-com links, so I removed b-com links & added11:57
digaohhh kk11:58
digaacabot: I will add that change11:58
digaacabot: sorry for misunderstanding11:58
acabotdiga : no pb12:00
digaacabot: pushed new PS -
digatox -edocs - successful12:26
alexchadinjed56: in this case we need to disable hypervisor 2 only after migration process for 5th VM is complete. I see here two ways: either to complete all the migration action, then complete all the disable action or to manage every subgraph independently. If we choose the second way, here is a question: where will we bind actions to each other? On the strategy layer or somehow on planner layer?12:27
jed56alexchadin: if I understand well12:31
jed56your question12:32
jed56we will bind or schedule the action in the planner12:32
jed56We can do the same thing12:32
jed56of your chart12:32
jed56I agree that the weight can't be set on the action in this case12:33
jed56The Watcher planner like the strategy are pluggable12:33
alexchadinjed56: how will planner know which actions should be binded?:)12:34
digaacabot: updated new PS -
jed56What do you mean by "binded" ?12:35
jed56I'm not sure to understand :)12:35
alexchadinjed56: Look - we have 5 VM and 2 hypervisors12:36
alexchadinWe want to disable the first hypervisor once 1-4 VMs are migrated12:36
alexchadinand we want to disable the second one once the 5th VM is migrated12:36
alexchadinso, how will we tell the system that disabling of the second hypervisor should come only after migrating of the 5th VMs?12:37
alexchadin5th VM*12:37
jed56So you question: How de we develop a heuristic that can schedule in parallel the migrations of the VM's per each hypervisor ?12:39
jed56that's it ?12:41
alexchadinNot exactly.12:41
jed56okay :p12:41
jed56Your question is regarding the way that we will store the results or the way that we will compute the order ?12:42
alexchadinMy question is: How can we define the vertexes of every subgraph?12:43
alexchadinjed56: yeah, how can we compute the order of vertexes!12:43
jed56okay :)12:44
jed56IMHO, if you have a parameter on the planner that limit the number of parallel migrations to 412:45
jed56you can select in the list actions the hypervisor with the largest number of VM's12:47
jed56create 4 vertex from the source node12:48
jed56then add a new node of level 2 add a new vertex12:48
alexchadinok, so we will launch actions in parallel by something like "layers" (we start with migrations, then continue with disabling node and so on). Right?12:48
alexchadindiga: ping12:49
jed56You can image many concurrent layer12:52
alexchadinjed56: so here we have at least two types of vertex - node and migration, right?12:53
alexchadinand disable12:53
alexchadinthe third:)12:53
sballe_jed56: it is a Holliday in the US today. Please send me your slides by EOD Wednesday and I will finalize them on Thu and Fri12:54
jed56The vertex is modeling a transition that can be concurrent if you have several vertex from one none12:55
alexchadinjed56: we start from level 0 and goes up?12:58
jed56You have to understand that we three notions13:00
jed56We have the node, the vertex and the task13:00
jed56In our case the task are the actions13:00
jed56For example in taskflow13:00
jed56r = gf.Flow("root")13:00
jed56n_1 = gf.Flow("NODE_1")13:01
jed56n_2 = gf.Flow("NODE_2")13:01
jed56class MigrateVM(task.Task):13:01
jed56    def execute(self):13:01
jed56        print("Running %s" %
jed56, n_1)13:02
jed56so NODE_1 and NODE_2 will be executed in concurrent13:03
jed56mig = MigrateVM(VM1)13:03
jed56In order to solve that we add new parameter13:04
jed56default planner = we can add a new parameter max 2 migration per hypervisor13:05
jed56and 10 migrations max in global13:05
jed56this is just an example but we need to  have the topology and the bandwidth available to compute a good solution13:05
alexchadinnow it makes more sense13:07
*** diga has quit IRC13:07
jed56Great ! :)13:07
alexchadinwell thank you a lot, now I understand what you want:)13:09
*** alexchadin has quit IRC13:09
*** alexchadin has joined #openstack-watcher13:09
jed56your welcome13:27
*** alexchadin has quit IRC13:31
*** hanrong has quit IRC13:49
*** michaelgugino has joined #openstack-watcher13:52
*** diga has joined #openstack-watcher14:12
digaacabot: vincentfrancoise : Hi14:12
diganeed your advise on alex's comment14:12
digaIn the previous he has proposed not to use audit/ & now he is again asking me to use audit/
digaacabot: I was also saying better to go with auto/, need your final call on this14:20
openstackgerritMykyta Karpin proposed openstack/puppet-watcher: Fix documentation for log_dir parameter
*** openstack has joined #openstack-watcher18:22
*** jed56 has quit IRC18:27
*** openstack has joined #openstack-watcher19:04
*** diga has quit IRC19:14
*** dtardivel has quit IRC19:24
*** michaelgugino has quit IRC20:23
*** thorst has quit IRC21:27
*** thorst has joined #openstack-watcher21:56
*** thorst has quit IRC22:01
*** thorst has joined #openstack-watcher23:10

Generated by 2.14.0 by Marius Gedminas - find it at!