Friday, 2017-02-10

*** thorst_ has joined #openstack-watcher01:03
*** thorst_ has quit IRC01:06
*** grumpycatt has quit IRC01:10
*** grumpycatt has joined #openstack-watcher01:11
*** yuanying has quit IRC01:12
*** yuanying has joined #openstack-watcher01:13
*** zhurong has joined #openstack-watcher01:13
*** thorst_ has joined #openstack-watcher01:37
*** thorst_ has quit IRC01:38
*** zhurong has quit IRC01:52
*** licanwei has quit IRC02:01
*** zhurong has joined #openstack-watcher02:12
*** zhurong has quit IRC02:25
*** zhurong has joined #openstack-watcher02:28
*** zhurong has quit IRC02:38
*** thorst_ has joined #openstack-watcher02:38
*** thorst_ has quit IRC02:48
*** thorst_ has joined #openstack-watcher02:48
*** zhurong has joined #openstack-watcher02:51
*** thorst_ has quit IRC02:53
*** harlowja has quit IRC02:53
*** zhurong has quit IRC02:59
*** thorst_ has joined #openstack-watcher03:09
*** thorst_ has quit IRC03:09
*** thorst_ has joined #openstack-watcher03:35
*** thorst_ has quit IRC03:35
*** adisky_ has joined #openstack-watcher04:42
openstackgerritOpenStack Proposal Bot proposed openstack/python-watcherclient master: Updated from global requirements
*** diga has joined #openstack-watcher05:01
*** thorst_ has joined #openstack-watcher05:36
*** thorst_ has quit IRC05:41
*** thorst_ has joined #openstack-watcher07:37
*** thorst_ has quit IRC07:42
*** hidekazu has joined #openstack-watcher08:03
*** vincentfrancoise has joined #openstack-watcher08:30
*** alexchadin has joined #openstack-watcher08:33
openstackgerritAlexander Chadin proposed openstack/watcher master: Add Action Notification
openstackgerritlicanwei proposed openstack/watcher-specs master:     Define when an action plan is stale/invalid
hidekazu'Define when an action plan is stale/invalid' is not implemented in ocata. but i am seeing SUPERSEDED everyday.09:28
vincentfrancoisehidekazu: hi09:29
hidekazuvincentfrancoise: hello09:29
vincentfrancoisehidekazu: the SUPERSEDED stated already exist but this BP makes tackles other use cases where we should also tag an action plan as superseded09:30
hidekazuvincentfrancoise: are there any bp implemented in ocata about SUPERSEDED/09:31
vincentfrancoisehidekazu: this one what meant to go in ocata but didn't make it09:31
hidekazuvincentfrancoise: as far as i understand, SUPERSEDED is first introduced in ocata, right?09:33
vincentfrancoisehidekazu: other that this one, was the one that introduced the SUPERSEDED state09:33
vincentfrancoisehidekazu: and this one is in ocata09:35
hidekazuvincentfrancoise: oh, auto trigger includes SUPERSEDED..09:35
hidekazuvincentfrancoise: thanks, i understand now.09:36
vincentfrancoisehidekazu: the thing that makes it a bit confusing is that we thought both BPs could be merged and only 1 made it into ocata09:37
vincentfrancoisehidekazu: which is why we ended up not having a clear reference to this new SUPERSEDED state09:37
*** thorst_ has joined #openstack-watcher09:38
hidekazuvincentfrancoise: yes even we contributer misunderstand ;)09:38
vincentfrancoisehidekazu: the doc has been updated as well09:39
vincentfrancoisehidekazu: yeah it's difficult to follow everything09:39
hidekazuvincentfrancoise: i checked today.09:40
*** thorst_ has quit IRC09:43
*** dtardivel has joined #openstack-watcher10:05
*** openstackgerrit has quit IRC10:18
*** hidekazu has quit IRC11:06
*** alexchadin has quit IRC11:12
*** vincentfrancoise has quit IRC11:25
*** diga has quit IRC11:31
*** alexchadin has joined #openstack-watcher11:31
*** thorst_ has joined #openstack-watcher11:39
*** openstackgerrit has joined #openstack-watcher11:43
openstackgerritaditi sharma proposed openstack/watcher-specs master: specs for blueprint stop action plan
*** thorst_ has quit IRC11:43
*** ri0 has joined #openstack-watcher11:54
*** thorst_ has joined #openstack-watcher12:33
*** ri0 has quit IRC12:39
*** alexchadin has quit IRC12:52
*** vmahe has joined #openstack-watcher13:09
*** vmahe has left #openstack-watcher13:09
*** vincentfrancoise has joined #openstack-watcher13:20
pshedimbvincentfrancoise, hi14:07
pshedimbI am working on noisy neighbor strategy and started developing some code based on our basic algorithm14:07
vincentfrancoisepshedimb: hi14:07
pshedimbI have a question to you14:08
vincentfrancoisepshedimb: great! please go ahead14:08
pshedimb in decision_engine/model/, there is a method get_node_instances14:09
*** adisky_ has quit IRC14:09
pshedimbWhen I use this method, the list the comes from self.predecessors(node_uuid) is empty14:09
pshedimbI tried debugging exisitng strategies and it comes empty for them as well14:10
pshedimbWhen I seperately call _get_by_uuid(instance_uuid), I get all the details of instance14:10
pshedimbWhat do you think the issue is?14:11
pshedimbAny mapping issue from node to instance?14:11
vincentfrancoisepshedimb: the call should be self.predecessors(node.uuid) instead of self.predecessors(node) I think14:12
pshedimbYeah, this is the exact statement14:14
pshedimbfor instance_uuid in self.predecessors(node.uuid):14:14
pshedimbSo self.predecessors(node.uuid) gives me empty list.14:14
vincentfrancoisepshedimb: give me 30mn so I can do some tests on this14:15
openstackgerritMerged openstack/watcher master: Fix the mapping between the instance and the node
pshedimbOkay thank you so much vincentfrancoise14:16
vincentfrancoisepshedimb: can you send me the call you make that doesn't work as well please?14:19
pshedimbvincentfrancoise, For debuggin this, I modified strategies/outlet_temp.py14:21
pshedimbThis is what I added.14:21
pshedimbfor node in nodes.values():14:21
pshedimb            resource_id = node.uuid14:21
pshedimb            import pdb; pdb.set_trace()14:21
pshedimb            instances_of_node = self.compute_model.get_node_instances(node)14:21
pshedimbI added a debug point here as well14:23
vincentfrancoisepshedimb: ok let me test this out14:23
pshedimbok thank you. I'm starting to drive to work (long drive), but will keep looking out for your messages and trying to reply asap14:25
vincentfrancoisepshedimb: hum14:27
vincentfrancoiseget_node_instances takes a ComputeNode object as parameter14:27
vincentfrancoiseand we gave it a string14:27
pshedimbvincentfrancoise, I passed ComputeNode object itself14:32
pshedimbvincentfrancoise, here is the code where I pass ComputeNode object to get_node_instances14:35
vincentfrancoisepshedimb: are you trying this via unit test or on a real infra?14:37
vincentfrancoisepshedimb: because just got merged and it fixes some mapping issue14:38
pshedimbvincentfrancoise, On real infra14:38
vincentfrancoisepshedimb: can you retry after pulling this change?14:39
pshedimbvincentfrancoise, okay. I'll test it out after pulling the chnages and will let you know14:39
pshedimbthank you :)14:39
vincentfrancoiseok I'll be available for the next 2h or so14:40
pshedimbWill pull the change now. will get back to you within 2 hours14:43
vincentfrancoisepshedimb: thanks14:43
*** Yumeng has left #openstack-watcher15:16
*** michaelgugino has joined #openstack-watcher15:36
*** edleafe is now known as figleaf15:37
*** cfriesen has joined #openstack-watcher15:38
cfriesengiven that nova has stated they want to get out of the metrics gathering business...can someone point me at the options for gathering compute node metrics to feed into watcher?15:39
vincentfrancoisecfriesen: hi15:41
vincentfrancoisecfriesen: I didn't know nova wanted to do that: do you have the mail/reference where they stated this please?15:41
vincentfrancoisecfriesen: Watcher relies on either Monasca or Ceilometer(now deprecated) to get the metrics15:42
vincentfrancoisecfriesen: so the question for Watcher is whether or not Monasca will collect the compute node metrics themselves from now on15:42
vincentfrancoisecfriesen: also, Watcher can be plugged to any metrics collection backend you want really, so if you want to deploy something else it's up to you (see
vincentfrancoisecfriesen: FYI, we also plan on supporting Gnocchi by the end of Pike15:46
cfriesenvincentfrancoise:   also see  about 2/3 down at, "Future of compute host metrics monitoring in Nova"15:47
cfriesenbasically they're not adding more metrics gathering in nova15:48
vincentfrancoisecfriesen: thanks15:50
vincentfrancoisecfriesen: we don't collect metrics from Nova directly but through the Ceilometer API15:51
vincentfrancoisecfriesen: but with the deprecation of Ceilometer API, this problem will disappear with it.15:51
vincentfrancoisecfriesen: did I fully answer your question?15:54
*** jwcroppe has joined #openstack-watcher16:01
pshedimbvincentfrancoise, after the new changes I pulled, it works fine :)16:05
vincentfrancoisepshedimb: cool :)16:05
cfriesenvincentfrancoise: yes, that's helpful.16:05
*** cfriesen has left #openstack-watcher16:05
pshedimbthanks vincentfrancoise. I can continue my development. :)16:06
*** acabot has quit IRC16:09
*** vincentfrancoise has quit IRC16:48
*** dtardivel has quit IRC17:00
*** harlowja has joined #openstack-watcher17:05
*** harlowja has quit IRC17:05
*** harlowja has joined #openstack-watcher17:05
*** wootehfoot has joined #openstack-watcher17:08
*** wootehfoot has quit IRC18:04
*** wootehfoot has joined #openstack-watcher18:12
*** thorst_ has quit IRC22:09
*** thorst_ has joined #openstack-watcher22:29
*** thorst_ has quit IRC22:34
*** michaelgugino has quit IRC22:43
*** figleaf is now known as edleafe23:13
*** jwcroppe has quit IRC23:34
*** jwcroppe has joined #openstack-watcher23:35
*** jwcroppe has quit IRC23:39

Generated by 2.14.0 by Marius Gedminas - find it at!