Monday, 2016-07-11

« Sunday, 2016-07-10 Index Tuesday, 2016-07-12 »
*** thorst has joined #openstack-watcher00:33
*** thorst has quit IRC00:35
*** thorst has joined #openstack-watcher00:36
*** esberglu has joined #openstack-watcher00:41
*** thorst has quit IRC00:44
*** esberglu has quit IRC01:21
*** thorst has joined #openstack-watcher01:27
*** thorst has quit IRC01:27
jinquanhello? anybody?01:37
*** esberglu has joined #openstack-watcher02:10
openstackgerritRong Han proposed openstack/watcher-dashboard: [WIP]Add policies for API access control to watcher-dashboard  https://review.openstack.org/33908202:47
openstackgerritRong Han proposed openstack/watcher-dashboard: [WIP]Add policies for API access control to watcher-dashboard  https://review.openstack.org/33908202:53
*** esberglu has quit IRC03:01
openstackgerritRong Han proposed openstack/watcher-dashboard: [WIP]Add policies for API access control to watcher-dashboard  https://review.openstack.org/33908205:36
*** acabot_ has joined #openstack-watcher06:49
acabotjinquan : hi06:50
*** acabot_ has quit IRC06:57
*** acabot_ has joined #openstack-watcher07:00
*** alexchadin has joined #openstack-watcher07:05
*** diga has joined #openstack-watcher07:13
*** acabot_ has quit IRC07:21
alexchadinhi acabot07:22
*** dtardivel has joined #openstack-watcher07:30
alexchadinhi gzhai207:33
*** diga has quit IRC07:35
jinquanacabot: hi07:46
jinquanguys:hi07:46
alexchadinhi jinquan07:47
jinquani commit  a patch set without a bug.  the patch set already merge.  now someone advice me that add a bug in the patch set.07:48
jinquando you know how to add a bug in a patch set which already merge?07:48
jinquanalexchadin: hi:)07:48
alexchadinjinquan: do you want to amend your merged commit?07:49
*** acabot_ has joined #openstack-watcher07:49
*** vincentfrancoise has joined #openstack-watcher07:49
jinquanyes , amend the commit msg07:51
alexchadinhm07:52
alexchadinIf it is in closed state, I think you can't amend it07:52
*** danpawlik has joined #openstack-watcher07:53
jinquanalexchadin: yes, i agree, but what can i do now?07:54
*** alexchadin has quit IRC07:54
danpawlikgood morning07:55
*** alexchadin has joined #openstack-watcher07:55
gzhai2vincentfrancoise: Morning07:56
openstackgerritMerged openstack/watcher: Uniform Airflow migration strategy implementation  https://review.openstack.org/32416907:56
vincentfrancoisegzhai2: morning08:00
gzhai2vincentfrancoise: wanna discuss with you for the sync issue.08:02
vincentfrancoisewhat sync issue?08:03
gzhai2multiple decision engine starting sync at the same time cause race condition.08:03
gzhai2https://bugs.launchpad.net/watcher/+bug/159894608:03
openstackLaunchpad bug 1598946 in watcher "Initial execution of parallel engines can be racey " [Medium,In progress] - Assigned to Zhai, Edwin (edwin-zhai)08:03
gzhai2this is a bug just FYI: https://review.openstack.org/#/c/339285/08:03
gzhai2We need make each decision engine is aware of each other when sync.08:04
*** diga has joined #openstack-watcher08:05
gzhai2One tricky issue is: Engine A just create goal A, then engine B treat it as stale and remove it immediately. So Engine A create stratey based on goal A failed.08:05
gzhai2Currently, we use schema entry point as the schema of efficacy specification in goal, so each engine treates the same goal as different.08:07
gzhai2vincentfrancoise: Just wanna some suggestion from you. I think we need hash the schema, instead of the entry point of function iteself.08:17
gzhai2e.g one schema of indicator specification is:         return voluptuous.Schema(08:19
*** alexchadin has quit IRC08:19
gzhai2            voluptuous.Range(min=0, max=100), required=True)08:19
*** alexchadin has joined #openstack-watcher08:20
*** jinquan has quit IRC08:20
*** jinquan has joined #openstack-watcher08:21
gzhai2the source code of schema remains same even the min/max changed. So I'd like define a string for the schema, which is used for detect changing of the schema.08:25
vincentfrancoisesorry I'm back08:27
vincentfrancoiseAh I see08:29
vincentfrancoiseThere are multiple issues about deploying multiple decision engine processes08:31
vincentfrancoisesince some of its processing is pretty much stateful08:33
vincentfrancoisebut clearly that's something we should improve08:35
vincentfrancoiseIMHO, this problem requires a BP because changing the way we sync the model may be structuring08:36
openstackgerritRong Han proposed openstack/watcher-dashboard: [WIP]Add policies for API access control to watcher-dashboard  https://review.openstack.org/33908208:52
*** alexchadin has quit IRC08:52
*** bauwser is now known as bauzas09:00
*** hanrong1 has joined #openstack-watcher09:01
*** alexchadin has joined #openstack-watcher09:04
alexchadinhi vincentfrancoise09:05
vincentfrancoisealexchadin: hi09:05
alexchadinI think ResourceType should be bound to the Scope Resources09:06
alexchadinRemember my latest spec09:06
alexchadin?09:06
alexchadinhttps://review.openstack.org/#/c/339456/1/specs/newton/approved/define-the-audit-scope.rst09:07
vincentfrancoiseWhat do you mean by "scope resources" ?09:08
alexchadinNovaScopeResource, CinderScopeResource classes09:08
alexchadinwhere we can specify resource types09:08
vincentfrancoiseah yeah09:08
vincentfrancoisethe problem is that you have 2 moving parts09:09
vincentfrancoise1 is the structure with instances, compute nodes and later things like volumes, ...09:09
vincentfrancoisethese resources do not require us to go through nova really09:10
vincentfrancoisebut the resources types need it if I'm not mistaken09:10
alexchadinWe have model that contains resources (cpu, ram, etc). We could insert one more level between model and resource and call it service09:11
alexchadinand when CDMC working in background09:12
alexchadinIt will call every ScopeResource to get latest info about resources it have09:13
*** jed56 has joined #openstack-watcher09:14
alexchadinso we could use something like model.cinder.volume_mb.get_capacity_from_id(cinder_volume_uuid)09:16
vincentfrancoisehmm, scoped resources are only relevant when you run a strategy and someone specified an audit scope09:17
vincentfrancoiseso the CDMC does not really have to be aware of the scope when running in the background09:18
alexchadinok09:31
alexchadinwhat about inserting service level?09:32
alexchadinwe could use it to create subset of cluster model when strategy is preparing to start09:32
alexchadinvincentfrancoise:09:32
vincentfrancoisealexchadin: sorry, I was talking about this issue with jed56 since he was leaving09:37
vincentfrancoisejust FYI, he agrees there is an overlap between you BP and mine regarding the structure of a the CDM structure09:38
vincentfrancoisewhat he had in mind would be a complete rework of the structure but he didn't have much time to lay it down to me09:40
alexchadinlet me write my vision of this issue09:40
vincentfrancoiseI am currently suggesting to him an XML representation of our model09:41
vincentfrancoisewhere we could query things using xpath to update it fairly easily09:41
vincentfrancoisealso you could use some XLST to perform your filtering09:41
vincentfrancoisejed56's solution would be based on Ecore (http://download.eclipse.org/modeling/emf/emf/javadoc/2.9.0/org/eclipse/emf/ecore/package-summary.html#details)09:44
alexchadinI'll be back in hour09:45
alexchadinlunch time09:45
*** alexchadin has quit IRC09:46
openstackgerritMerged openstack/python-watcherclient: Prints  '-' instead of 'None' when data is None  https://review.openstack.org/34013910:00
*** alexchadin has joined #openstack-watcher10:33
*** thorst has joined #openstack-watcher10:46
*** thorst has quit IRC10:47
*** thorst has joined #openstack-watcher10:47
*** thorst_ has joined #openstack-watcher10:50
*** thorst has quit IRC10:52
alexchadinvincentfrancoise: have you get any solution with jed56?10:56
alexchadingot*10:56
*** jed56 has quit IRC11:35
vincentfrancoisealexchadin: well, only what I told before lunch break11:45
vincentfrancoisebut he told he'll spend some time with me on wednesday we he gets back11:46
vincentfrancoisein the meatime I'll try to get my XML-based structure to work to see if it's enough for now11:47
vincentfrancoisewhat about you?11:47
alexchadinI will try to write something like spec where describe my thoughts about module structure of resources11:48
alexchadinand update continuous mechanism with gzhai2 suggestions and job storing11:49
alexchadinis it good for you?11:50
vincentfrancoiseabout resources, I'm afraid one of us will have to throw away its solution11:52
vincentfrancoiseso do not write a full spec, just the necessary to understand what you suggest11:53
vincentfrancoiseI will do kind of the same although I know jed56 may propose something different as well11:54
vincentfrancoiseit's actually a big/critical subject...11:54
alexchadinvincentfrancoise: yeah, it would be interesting to know his propositions12:02
digaalexchadin: Hi12:15
alexchadinhi diga12:16
digaalexchadin: I am writting spec on the assigned BP12:16
digaalexchadin: I am going through your BP & spec12:16
digahow I can test your code, is it in progress still ?12:17
alexchadinIt was merged already12:18
digaokay12:18
digaI will go through that12:18
alexchadinBut CLI and Dashboard are on the way to get merged12:18
digaok12:18
alexchadinonce CLI is merged you can check continuous type using options -t CONTINUOUS -i 120 as example12:19
digaalexchadin: your location ?12:19
digaokay12:19
alexchadindiga: location?12:19
digaregion, country etc12:20
digacity12:20
digaI am from Pune, India12:20
alexchadinah, sorry, Moscow, +3 UTC12:20
digabecause I am leaving office now, will resume from home around 7:30 PM my time12:21
digaso just checking whether you will be available till then or not12:21
alexchadinI will be available during 1.5 hours12:22
digaokay12:22
alexchadinIf you want, you can write email12:22
digaokay12:22
digaemail id ?12:23
alexchadinor leave your comment in gerrit topic/launchpad12:23
digaokay12:23
digasure12:23
digathanks alexchadin12:23
*** diga has quit IRC12:30
*** alexchadin has quit IRC13:30
*** alexchadin has joined #openstack-watcher13:30
*** esberglu has joined #openstack-watcher13:39
*** alexchadin has quit IRC13:48
sballe_Morning13:49
sballe_acabot: please see: https://review.openstack.org/#/c/329873/1413:51
sballe_Is the meeting with the nova team today?13:51
vincentfrancoisesballe_: hi14:08
vincentfrancoisesballe_: acabot is not there at the moment14:08
sballe_ok thx :-)14:08
vincentfrancoisebut I believe the nova meeting it is right now14:08
vincentfrancoiseon #openstack-meeting-alt14:08
vincentfrancoisehere is their agenda https://wiki.openstack.org/wiki/Meetings/NovaScheduler14:09
*** diga_ has joined #openstack-watcher14:21
sballe_hanrong1: what's your email for talk submission?14:25
sballe_vincentfrancoise: please update your name and email on https://etherpad.openstack.org/p/Watcher_abstracts_barcelona201614:27
vincentfrancoisesballe_: ok14:31
openstackgerritMerged openstack/puppet-watcher: Include openstacklib defaults manifest.  https://review.openstack.org/33884814:33
*** brunograz has joined #openstack-watcher14:46
*** alexchadin has joined #openstack-watcher14:49
*** alexchadin has quit IRC14:52
*** hvprash has joined #openstack-watcher14:55
*** alexchadin has joined #openstack-watcher15:22
*** alexchadin has quit IRC15:27
acabotsballe_ : I'm reviewing it15:30
sballe_thx15:30
acabotsballe_ : I tried to push it in the nova scheduler meeting :-)15:30
acabotsballe_ : I tried....15:30
sballe_so they didn;t discuss it15:30
sballe_when is there next meeting?15:31
acabothttp://eavesdrop.openstack.org/meetings/nova_scheduler/2016/nova_scheduler.2016-07-11-14.00.log.html15:31
sballe_ok thx15:31
acabotthey suggest to have it discussed during an unconference session15:31
sballe_ok but isn't that too far out for us?15:31
acabotthe problem is that there is now unconference session at the midcycle :-)15:31
acabots/now/no15:32
sballe_can we discuss it at the mid-cycle?15:32
acabotso I will add it to the Nova agenda15:32
acabotbut I dont know when it will be discussed...mriedem has too many things in the backlog and dont want to discuss things not related to Newton15:33
*** danpawlik has quit IRC15:45
*** brunograz has quit IRC15:50
*** alexchadin has joined #openstack-watcher16:28
*** alexchadin has quit IRC16:33
diga_acabot: Hi16:38
diga_acabot: I have some questions on watcher event collection ?16:39
*** vincentfrancoise has quit IRC16:45
*** wootehfoot has joined #openstack-watcher17:49
*** wootehfoot has quit IRC18:05
*** wootehfoot has joined #openstack-watcher18:10
*** wootehfoot has quit IRC18:35
*** wootehfoot has joined #openstack-watcher18:36
*** alexchadin has joined #openstack-watcher19:01
*** alexchadin has quit IRC19:02
openstackgerritDaniel Pawlik proposed openstack/puppet-watcher: Implement watcher-db-manage commands  https://review.openstack.org/33693919:22
openstackgerritDaniel Pawlik proposed openstack/puppet-watcher: Implement watcher-db-manage commands  https://review.openstack.org/33693919:32
*** dtardivel has quit IRC19:37
*** wootehfoot has quit IRC20:16
openstackgerritMerged openstack/puppet-watcher: Implement watcher-db-manage commands  https://review.openstack.org/33693920:51
*** thorst_ has quit IRC20:56
*** thorst_ has joined #openstack-watcher20:57
*** thorst_ has quit IRC21:02
*** alexchadin has joined #openstack-watcher21:02
*** thorst_ has joined #openstack-watcher21:03
*** alexchadin has quit IRC21:05
*** thorst_ has quit IRC21:07
openstackgerritDaniel Pawlik proposed openstack/puppet-watcher: Implement applier and decision-engine  https://review.openstack.org/33693521:18
*** hvprash has quit IRC21:28
*** thorst_ has joined #openstack-watcher21:52
*** esberglu has quit IRC21:55
*** thorst_ has quit IRC21:57
*** thorst has joined #openstack-watcher22:31
*** thorst has quit IRC22:34
*** thorst has joined #openstack-watcher22:35
*** thorst has quit IRC22:39
« Sunday, 2016-07-10 Index Tuesday, 2016-07-12 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!