Monday, 2016-06-06

*** thorst has joined #openstack-watcher00:55
*** thorst has quit IRC00:56
*** Kevin_Zheng has joined #openstack-watcher01:43
*** thorst has joined #openstack-watcher01:56
*** thorst has quit IRC01:59
*** thorst has joined #openstack-watcher01:59
*** thorst has quit IRC02:00
*** thorst has joined #openstack-watcher02:00
*** thorst has quit IRC02:05
*** esberglu has joined #openstack-watcher02:39
*** esberglu has quit IRC02:41
*** thorst has joined #openstack-watcher03:03
*** thorst has quit IRC03:10
*** thorst has joined #openstack-watcher04:09
*** thorst has quit IRC04:16
*** jwcroppe has joined #openstack-watcher04:46
*** jwcroppe has quit IRC04:46
*** jwcroppe has joined #openstack-watcher04:47
*** jwcroppe has quit IRC05:04
*** jwcroppe has joined #openstack-watcher05:10
*** thorst has joined #openstack-watcher05:14
*** ecelik has joined #openstack-watcher05:17
*** thorst has quit IRC05:20
*** openstackgerrit has quit IRC06:02
*** openstackgerrit has joined #openstack-watcher06:03
*** thorst has joined #openstack-watcher06:18
*** wootehfoot has joined #openstack-watcher06:22
*** thorst has quit IRC06:26
*** vmahe has joined #openstack-watcher06:33
*** apoorv has joined #openstack-watcher06:37
*** jwcroppe has quit IRC06:54
*** jwcroppe has joined #openstack-watcher06:56
*** jwcroppe has quit IRC07:00
*** jed56 has joined #openstack-watcher07:22
*** thorst has joined #openstack-watcher07:25
*** thorst has quit IRC07:31
*** vtech has joined #openstack-watcher07:31
*** vincentfrancoise has joined #openstack-watcher07:34
*** alexchadin has joined #openstack-watcher08:14
*** wootehfoot has quit IRC08:22
*** thorst has joined #openstack-watcher08:29
*** thorst has quit IRC08:35
*** apoorv has quit IRC08:42
*** apoorv has joined #openstack-watcher08:51
*** alexchadin has quit IRC08:53
*** alexchadin has joined #openstack-watcher08:53
*** danpawlik has quit IRC08:54
*** danpawlik has joined #openstack-watcher08:54
*** alexchadin has quit IRC08:57
*** alexchadin has joined #openstack-watcher09:00
vincentfrancoisehi gzhai209:33
*** thorst has joined #openstack-watcher09:33
*** thorst has quit IRC09:40
openstackgerritVincent Françoise proposed openstack/watcher: Added filter operators
*** wootehfoot has joined #openstack-watcher09:46
*** wootehfoot has quit IRC09:46
openstackgerritVincent Françoise proposed openstack/watcher: Added filter operators
*** alexchadin has quit IRC09:54
*** alexchadin has joined #openstack-watcher09:54
*** thorst has joined #openstack-watcher10:38
*** alexchadin has quit IRC10:38
*** alexchadin has joined #openstack-watcher10:39
*** alexchadin has quit IRC10:43
*** alexchadin has joined #openstack-watcher10:43
*** thorst has quit IRC10:46
*** alexchadin has quit IRC10:51
*** alexchadin has joined #openstack-watcher11:35
*** thorst has joined #openstack-watcher11:36
*** alexchadin has quit IRC11:38
*** apoorv has quit IRC12:19
*** hvprash has joined #openstack-watcher12:30
*** hvprash has quit IRC12:31
openstackgerritAntoine Cabot proposed openstack/watcher: Add reno for release notes management
vincentfrancoisegzhai2: hi12:58
gzhai2about my patch, want talk with you.12:59
*** figleaf is now known as edleafe12:59
gzhai2about jsonschema. I can use it, but the issue is we need do the validation in strategy code rather than API13:00
gzhai2As API has no knowledge of which strategy is used, so can't locate the right schema.13:01
vincentfrancoiseyes but the persistent-audit-parameters blueprint should solves this issue (
vincentfrancoiseso maybe there is a missing dependency then13:02
vincentfrancoiseso maybe you can add the strategy_id foreign key to the Audit model so you can tell which strategy the audit is applying in the meantime13:04
gzhai2Then let's merge optimization-threshold first, then change it after persistent-audit-parameters.13:05
gzhai2I can add foreign key, but it's only be done in audit creation, where we alredy know the strategy id.13:06
vincentfrancoiseah yeah as I said in my comment, your audit doesn't know about the strategy if you are in "case 1"13:07
vincentfrancoisebut in case 2, your audit knows which strategy it applies since it's the same as the related audit template13:08
gzhai2What's case 1 and 2?13:10
vincentfrancoise ->Case 1: If no strategy was provided, I agree we cannot   validate the parameters: should we still be able to   provide parameters other than the defaults if don't even   know what shall be run? BTW, that is a case we have yet   to discuss as it is not described in the specs. ->Case 2: If a strategy has been provided, then we know   the parameters we should apply, which means we can   potentially validate the strategy parameters at t13:10
vincentfrancoise(copy/paste from
gzhai2You want to different ways to handle case 1 & 2? The way in my patch for 1, while validation at API for 2?13:14
vincentfrancoisegzhai2: I'm not sure we should allow audit parameters unless we know in advance which strategy we are targeting13:15
vincentfrancoisegzhai2: that's just my opinion so we should probably ask the others about it13:16
openstackgerritVincent Françoise proposed openstack/watcher: Added efficacy indicators to /action_plans
openstackgerritVincent Françoise proposed openstack/watcher: [WIP] Strategy selector select based on efficacy
openstackgerritVincent Françoise proposed openstack/watcher: Documentation on goal and efficacy
openstackgerritVincent Françoise proposed openstack/watcher: Added pre/post execution methods to strategies
openstackgerritVincent Françoise proposed openstack/watcher: [WIP] Cancel out related audits and action plans
gzhai2How about raise exception if user input parameter for non-parameter strategy?13:17
gzhai2You know. flexibilty is most important for strategy selection. I'm not sure if we can know which strategy would be used.13:18
vincentfrancoiseWe may do this but IMHO this is misleading: if you give some parameters, this means you know which strategy you want to use13:18
gzhai2Especially in long term.13:18
vincentfrancoiseotherwise, this means that some parameters should be shared because of the goal13:19
gzhai2My original patch instantiate one strategy when creating audit-template to get the parameter info. Is it ok?13:22
gzhai2Currently, user need search the strategy list to find the goal-strategy mapping.13:22
gzhai2My point is: strategy-goal solution is not clear at this point, like how to map multiple stratey to one goal. I believe we need some enhancement in future.13:24
jed56hello gzhai213:25
gzhai2But strategy parameter is critical feature for watcher usage, so I want a for-now solution and change it in future.13:25
jed56Exactly, we add a feature to give to watcher the ability to select the best strategy13:25
gzhai2jed56: hi13:25
jed56You can take a look to this blueprint
*** hvprash has joined #openstack-watcher13:26
jed56This is the reason why, if don't "force" a strategy you can't specify the parameters at this moment13:26
gzhai2I know we have a lot of features to make Watcher mature, and I can wait for other features merged. But this is also critical features asked from last release.13:28
jed56yes I agree.13:29
jed56you can implement the parameters13:29
jed56you just can't specify parameter if you don't specify a strategy13:30
gzhai2This depends on our development strategy: long time perfect solution, or workable solution for now and change it gradually in future.13:30
jed56gzhai2: I agree, we have to build watcher incrementally13:31
jed56I don't understand well your point ? :)13:31
vincentfrancoisegzhai2: FYI, how can we generate the form on the watcher-dashboard to fill the strategy parameters if we don't know which one we want to apply?13:31
gzhai2How about instantiate one strategy when audit tmplate creation to know the right stratey?13:32
*** jwcroppe has joined #openstack-watcher13:32
*** hvprash_ has joined #openstack-watcher13:32
gzhai2Curently, I assume one goal only have one strategy, so user can search strategy list to find the right strategy for the goal.13:33
gzhai2If it's not ture, I'd like to add extra instantiating strategy for check.13:34
*** hvprash has quit IRC13:35
jed56IMHO, you should keep the code simple. If audit_template has not strategy you can't specify parameters13:36
gzhai2Then I can add strategy foreign key for audit-template.13:36
jed56otherwise, if the audit_template have a strategy you can13:36
vincentfrancoisegzhai2: one goal (server_consolidation) already has 2 strategies related to it (basic and vm_workload_consolidation), so cannot assume that13:37
gzhai2jed56: ok13:37
gzhai2You guys still prefer jsonschema, and validation in API?13:38
jed56ghzai do you have an alternative ?13:38
gzhai2If keep it as simple as you said, it's ok.13:40
gzhai2ok. I'll change it.13:41
gzhai2Thanks for your comments.13:41
jed56gzhai2:  you are welcome.13:42
vincentfrancoisethanks for making the changes ;)13:43
jed56if need help don't hesitate13:43
*** brunograz has joined #openstack-watcher13:48
brunograzHi guys13:48
brunograzWe have a testbed with watcher installed on it (we did the installation some months ago and actually never updated it)13:49
brunograzI did the update to the latest version of watcher13:49
brunograzand we are running into some issues - watcherclient has changed as well as the base strategy13:50
vincentfrancoisehi brunograz13:50
brunograzI had to drop the database and recreate it in order to use the new client13:50
brunograzbut now there are no goals inserted13:51
vincentfrancoiseyes it changed with the n1 release13:51
vincentfrancoisegoals are now defined via the strategy itself13:51
*** ecelik has quit IRC13:51
vincentfrancoiseand the decision engine syncs the goals and strategies with the DB to let the API expose it13:52
brunograzbut how do I insert the goal into the DB?13:53
brunograzthere is no option to create it and in my strategy I added the classmethod to return the goal name13:53
vincentfrancoisethe decision engine does it automatically when you start it13:53
brunograzhmm got it13:54
brunograzlet me try to restart it13:54
brunograzyep it worked (y)13:54
brunograzthx a mil13:55
brunograzmy error was to restart the decision engine before recreating the database13:55
vincentfrancoiseoh I see13:55
vincentfrancoiseI guess we will have to start doing DB migrations since we are now into the big tent13:56
vincentfrancoisemaybe we can put this in the open discussion of this week's meeting13:56
brunograzand I think we will have to file another bug to our strategy13:59
brunograzand it does not implement the new abstract methos13:59
brunograzopss sorry - it is implemeted14:01
*** esberglu has joined #openstack-watcher14:01
vincentfrancoiseit should be there otherwise some tests would have failed ;)14:05
*** hvprash has joined #openstack-watcher14:28
*** hvprash_ has quit IRC14:32
*** jinquan has joined #openstack-watcher14:37
*** jwcroppe has quit IRC14:55
*** jwcroppe has joined #openstack-watcher14:56
*** jwcroppe has quit IRC15:01
*** alexchadin has joined #openstack-watcher15:05
*** jinquan has quit IRC15:06
vincentfrancoisealexchadin: I reviewed you changeset on continuously-optimization15:09
alexchadinI think we can change audit type via API. For example, we have audit with some parameters. If we want to change continuous type to oneshot we don't need to create new audit with the same parameters15:10
alexchadinvincentfrancoise: I saw, thank you:)15:10
vincentfrancoisealexchadin: the problem is more about the fact that an audit is triggered as soon as it gets created. So even if you change the type of the audit, you shouldn't be able the reverse the state of the audit which will be already ONGOING or even SUCCEEDED15:12
alexchadinvincentfrancoise: let me think of it15:13
alexchadinvincentfrancoise: I agree15:15
danpawlikhello! I see, that in doc file in section: keystone_authtoken there are memcache* parameters, but I guess this memcache* parameters should be moved to cache section (according to puppet-oslo module - or?15:17
vincentfrancoisealexchadin: if you want to make sure you do not have to specify parameters over and over again, maybe we'll need to move these parameters up in the audit template as well15:17
alexchadinvincentfrancoise: I think having parameters in audit is well15:18
vincentfrancoisedanpawlik: Hi15:19
vincentfrancoisealexchadin: what we are talking about is pretty much what gzhai2 is working on BTW15:19
vincentfrancoisealexchadin: see
alexchadinvincentfrancoise: I know, I really want this changes to be merged ASAP:_15:21
vincentfrancoisedanpawlik: Agreed, we are missing "namespace = oslo.cache" in etc/watcher/watcher-config-generator.conf15:21
vincentfrancoisedanpawlik: but I don't know if memcache* params in keystone_authtoken are actually the same15:22
danpawlikvincentfrancoise: thanks. So I must confirm this params with guys from oslo/keystone/puppet etc. Hmm, maybe better if I make small changes in my puppet-watcher module and then community will say, what should I do ;P15:23
alexchadinvincentfrancoise: could you please give me an example of "IN" operator usage in your new PS?15:24
vincentfrancoisedanpawlik: give me 2 secs, I'll see if I can find anything in keystone_authtoken that would point to oslo.cache15:24
danpawlikvincentfrancoise: ok15:25
*** hvprash_ has joined #openstack-watcher15:25
vincentfrancoisealexchadin: line 17615:26
vincentfrancoisealexchadin: the docstring is a bit off by the way so I'll update it whenever I get some time15:28
alexchadinvincentfrancoise: it's okay, thank you!15:28
*** hvprash has quit IRC15:28
vincentfrancoisedanpawlik: seems to me that these memcached* parameters are different from the olso.cache ones15:35
danpawlikvincentfrancoise: ok. Good to know. Thanks for checking ;)15:37
vincentfrancoiseso that mean we need both these parameters and the missing oslo.cache ones15:37
*** alexchadin has quit IRC15:44
*** alexchadin has joined #openstack-watcher15:46
openstackgerritVincent Françoise proposed openstack/watcher: Added missing config section for autogeneration
vincentfrancoisedanpawlik: see ^15:48
danpawlikvincentfrancoise: looks good :D15:49
vincentfrancoisedanpawlik: there were way more than just oslo.cache that was actually missing :p15:50
danpawlikvincentfrancoise: I think that I can still find something xD15:52
vincentfrancoisedanpawlik: that wouldn't surprise me :p The more you find the better anyway15:52
danpawlikvincentfrancoise: yeah :D15:54
*** jwcroppe has joined #openstack-watcher15:57
*** brunograz has quit IRC16:00
*** jwcroppe has quit IRC16:03
*** vtech has quit IRC16:04
danpawlikhave a nice day / evening ;)16:06
vincentfrancoisedanpawlik: see you ;)16:07
*** danpawlik has quit IRC16:07
*** jinquan has joined #openstack-watcher16:08
*** vincentfrancoise has quit IRC16:16
*** jinquan has left #openstack-watcher16:23
*** vtech has joined #openstack-watcher16:44
*** jwcroppe has joined #openstack-watcher16:59
*** jwcroppe has quit IRC17:04
*** esberglu has quit IRC17:05
*** alexchadin has quit IRC17:15
*** jwcroppe has joined #openstack-watcher17:18
*** vtech has quit IRC17:29
*** alexchadin has joined #openstack-watcher17:32
*** harlowja has joined #openstack-watcher17:48
openstackgerritAlexander Chadin proposed openstack/watcher: Add continuously optimization
*** esberglu has joined #openstack-watcher18:02
*** jed56 has quit IRC18:45
*** alexchadin has quit IRC18:47
*** vtech has joined #openstack-watcher19:00
*** alexchadin has joined #openstack-watcher20:35
*** hvprash_ has quit IRC21:01
*** thorst has quit IRC21:17
*** jwcroppe has quit IRC21:30
*** thorst has joined #openstack-watcher21:36
*** jwcroppe has joined #openstack-watcher21:39
*** thorst_ has joined #openstack-watcher21:39
*** thorst has quit IRC21:42
*** jwcroppe_ has joined #openstack-watcher21:43
*** jwcroppe has quit IRC21:44
*** thorst_ has quit IRC21:45
*** thorst has joined #openstack-watcher21:45
*** jwcroppe_ has quit IRC21:54
*** thorst has quit IRC21:55
*** jwcroppe has joined #openstack-watcher21:55
*** jwcroppe has quit IRC21:59
*** esberglu has quit IRC22:20
*** thorst has joined #openstack-watcher22:52
*** alexchadin has quit IRC22:53
*** thorst has quit IRC23:00
*** harlowja has quit IRC23:01
*** jwcroppe has joined #openstack-watcher23:13
*** jwcroppe has quit IRC23:16
openstackgerritMerged openstack/watcher: Updated from global requirements
*** jwcroppe has joined #openstack-watcher23:55
*** thorst has joined #openstack-watcher23:58

Generated by 2.14.0 by Marius Gedminas - find it at!