*** thorst_ has joined #openstack-watcher | 01:25 | |
*** thorst_ has quit IRC | 01:28 | |
*** thorst_ has joined #openstack-watcher | 02:02 | |
*** thorst_ has quit IRC | 02:02 | |
openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/puppet-watcher: Add basic structure for ReNo https://review.openstack.org/383538 | 03:44 |
---|---|---|
*** diga has joined #openstack-watcher | 04:15 | |
*** hvprash has quit IRC | 04:55 | |
*** kamtamtun has joined #openstack-watcher | 06:28 | |
*** hvprash has joined #openstack-watcher | 06:37 | |
*** yuanying_ has joined #openstack-watcher | 06:40 | |
*** yuanying has quit IRC | 06:44 | |
openstackgerrit | Merged openstack/watcher: Enable release notes translation https://review.openstack.org/383408 | 07:16 |
*** dtardivel has joined #openstack-watcher | 07:23 | |
*** yuanying_ has quit IRC | 07:44 | |
*** yuanying has joined #openstack-watcher | 08:22 | |
*** yuanying has quit IRC | 08:26 | |
*** yuanying has joined #openstack-watcher | 08:26 | |
*** yuanying_ has joined #openstack-watcher | 08:28 | |
*** yuanying has quit IRC | 08:32 | |
openstackgerrit | David TARDIVEL proposed openstack/watcher: Add strategy template doc https://review.openstack.org/382494 | 08:33 |
*** alexchadin has joined #openstack-watcher | 08:51 | |
openstackgerrit | Alexander Chadin proposed openstack/watcher: Add service supervisor https://review.openstack.org/371632 | 08:52 |
*** yuanying_ has quit IRC | 08:52 | |
alexchadin | dtardivel, vincentfrancoise: thank you for your snippet! | 08:52 |
*** yuanying has joined #openstack-watcher | 08:52 | |
alexchadin | although I didn't want to modify utils.py, but it works at least:) | 08:52 |
dtardivel | you're welcome. fyi, I found another solution yesterday, based on ArrowType, but arrow lib is not in global requirements for now | 08:53 |
openstackgerrit | Alexander Chadin proposed openstack/python-watcherclient: Add support for Audit Scope https://review.openstack.org/364787 | 09:15 |
openstackgerrit | Alexander Chadin proposed openstack/python-watcherclient: Add support for Audit Scope https://review.openstack.org/364787 | 09:30 |
vincentfrancoise | alexchadin: morning | 09:38 |
alexchadin | vincentfrancoise: morning | 09:38 |
vincentfrancoise | alexchadin: yeah I didn't like it much | 09:38 |
vincentfrancoise | but since it will be scrapped anyway with the OVO blueprint I'm doing | 09:39 |
vincentfrancoise | it doesn't matter much | 09:39 |
alexchadin | vincentfrancoise: no problem:) | 09:39 |
alexchadin | vincentfrancoise: I'd like to discuss with you https://blueprints.launchpad.net/watcher/+spec/planner-storage-action-plan | 09:42 |
vincentfrancoise | alexchadin: ok | 09:43 |
alexchadin | Have you read it? | 09:43 |
vincentfrancoise | alexchadin: didn't read it per se but I know what it talks about ;) | 09:43 |
alexchadin | vincentfrancoise: there is subgraphs with linked actions | 09:45 |
alexchadin | vincentfrancoise: looks like strategy developer should build action dependencies by himself | 09:45 |
alexchadin | vincentfrancoise: Do you agree? | 09:45 |
vincentfrancoise | alexchadin: actions are linked together by the planner | 09:45 |
vincentfrancoise | alexchadin: a strategy should propose a cluster state to reach with the associated set of actions | 09:46 |
alexchadin | vincentfrancoise: ok, but how can we specify that, for example, these two actions are bounded? | 09:46 |
vincentfrancoise | alexchadin: how they should get played should be handled by the planner | 09:47 |
vincentfrancoise | linked actions do not refer to any kind of affinity | 09:48 |
vincentfrancoise | linked actions are just referring to the sequencing of actions | 09:48 |
vincentfrancoise | as of now there isn't any notion of action affinity | 09:48 |
vincentfrancoise | alexchadin: does that makes sense to you? | 09:50 |
alexchadin | Now we have linked list of actions. Each action is run after previous is done. jed56 speaks that each independent subgraph can be run in parallel. As I get it, action plan can contain different subgraphs. I wonder how planner can build these subgraphs? | 09:51 |
vincentfrancoise | alexchadin: as I understand it, it is based solely on the charactestics of each actions. i.e. Does that action can be parallelized? If so, can we build concurrent action pipelines which could run in parrallel? | 09:55 |
alexchadin | woah | 09:56 |
vincentfrancoise | each one of these concurrent pipelines would then be represented into the action graph as separate branches | 09:56 |
alexchadin | yeap | 09:56 |
alexchadin | now it better makes sense to me | 09:56 |
alexchadin | hm | 09:57 |
alexchadin | we have action that puts compute node in disabled state | 09:58 |
alexchadin | it could be prerequisite for migration action | 09:59 |
*** jed56 has joined #openstack-watcher | 09:59 | |
jed56 | hi | 10:00 |
alexchadin | hi jed56 | 10:00 |
vincentfrancoise | alexchadin: off for lunch, will be back in 2 | 10:01 |
jed56 | vincentfrancoise: said to me that you have some questions regarding the action-plan-storage | 10:01 |
alexchadin | jed56: yes. For example, we have the following scenario - two compute nodes, each of them should be turned into disabled state. To do so we need to migrate all the vms from them to another hosts. | 10:03 |
jed56 | true | 10:03 |
jed56 | Did you take a look to the current implementation of the planner ? | 10:04 |
alexchadin | As I get it, planner should create two subgraphs, each one will contain Migrate->Migrate->Migrate->Turn into disabled state pipeline, true? | 10:05 |
alexchadin | I'm looking to it currently | 10:07 |
jed56 | I you take a look to | 10:08 |
jed56 | https://github.com/openstack/watcher/blob/master/watcher/decision_engine/planner/default.py#L42 | 10:08 |
jed56 | you will see that we are using the weight to order the actions | 10:09 |
jed56 | The we store the ordered list in the watcher database | 10:09 |
jed56 | like that | 10:09 |
jed56 | start - migrate -> migrate -> migrate -> migrate -> disable hypervisor | 10:09 |
jed56 | We would like to be able to store the actions | 10:09 |
jed56 | https://www.irccloud.com/pastebin/bNG3xw58/ | 10:10 |
jed56 | like that | 10:10 |
jed56 | In the applier we are using http://docs.openstack.org/developer/taskflow/patterns.html#module-taskflow.patterns.graph_flow | 10:10 |
alexchadin | then | 10:10 |
alexchadin | we have created two subgraphs with lists of vms | 10:11 |
alexchadin | how will we know where to add properly disable action? | 10:12 |
alexchadin | I've seen graph_flow already, good tool | 10:13 |
jed56 | IMHO, we need one graph for the first version | 10:13 |
jed56 | The subgraph can be used for "alternative paths" | 10:14 |
jed56 | In my previous example, we need two edge | 10:14 |
jed56 | but four vertex | 10:15 |
jed56 | each vertex represent a concurrent action | 10:15 |
alexchadin | ok | 10:15 |
alexchadin | how can we define alternative paths? | 10:16 |
jed56 | IMHO, the alternative path can be "activate" with the pre_condition or "decider" in the taskflow terminology | 10:17 |
jed56 | decider – A callback function that will be expected to decide at runtime whether v should be allowed to execute (or whether the execution of v should be ignored, and therefore not executed). It is expected to take as single keyword argument history which will be the execution results of all u decidable links that have v as a target. It is expected to | 10:17 |
jed56 | return a single boolean (True to allow v execution or False to not). | 10:17 |
jed56 | But this is my opinion, you have maybe a better idea :) | 10:18 |
jed56 | are you comming to barcelona ? | 10:19 |
alexchadin | jed56: yes, I'll be there | 10:20 |
jed56 | great ;) | 10:21 |
alexchadin | Will draw some schemes for now to make it more sense:) | 10:22 |
jed56 | perfect | 10:22 |
*** alexchadin has quit IRC | 10:32 | |
*** alexchadin has joined #openstack-watcher | 10:33 | |
*** alexchadin has quit IRC | 10:40 | |
sballe_ | Morning | 11:11 |
*** thorst_ has joined #openstack-watcher | 11:48 | |
*** thorst_ has joined #openstack-watcher | 11:49 | |
*** mestery has quit IRC | 11:53 | |
*** mestery has joined #openstack-watcher | 11:53 | |
acabot | diga : hi | 11:53 |
diga | acabot: hi | 12:25 |
diga | acabot: I have submitted the PS as per your comments | 12:26 |
acabot | diga : yes but you have missed my last comment | 12:26 |
acabot | https://review.openstack.org/#/c/342473/ | 12:26 |
acabot | regarding dead links | 12:26 |
diga | ohhh | 12:26 |
diga | got it, I will update NP | 12:27 |
acabot | thx | 12:27 |
diga | wc! | 12:27 |
*** jimbaker has quit IRC | 12:38 | |
*** jimbaker has joined #openstack-watcher | 12:38 | |
*** jimbaker has quit IRC | 12:38 | |
*** jimbaker has joined #openstack-watcher | 12:38 | |
openstackgerrit | Vincent Françoise proposed openstack/watcher-specs: Specs for watcher-versioned-objects https://review.openstack.org/383718 | 12:55 |
*** edleafe is now known as figleaf | 12:56 | |
openstackgerrit | Vincent Françoise proposed openstack/watcher-specs: Specs for watcher-versioned-objects https://review.openstack.org/383718 | 13:04 |
openstackgerrit | Vincent Françoise proposed openstack/watcher-specs: Specs for watcher-versioned-objects https://review.openstack.org/383718 | 13:21 |
openstackgerrit | Vincent Françoise proposed openstack/watcher-specs: Specs for watcher-versioned-objects https://review.openstack.org/383718 | 13:31 |
*** diga has quit IRC | 13:46 | |
*** michaelgugino has quit IRC | 14:00 | |
*** aspiers has quit IRC | 14:22 | |
*** sballe_ has quit IRC | 14:22 | |
*** aspiers has joined #openstack-watcher | 14:23 | |
*** sballe_ has joined #openstack-watcher | 14:26 | |
*** hvprash_ has joined #openstack-watcher | 14:28 | |
openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/puppet-watcher: Add basic structure for ReNo https://review.openstack.org/383538 | 15:01 |
*** michaelgugino has joined #openstack-watcher | 15:23 | |
*** michaelgugino has quit IRC | 15:33 | |
*** michaelgugino has joined #openstack-watcher | 15:40 | |
*** vincentfrancoise has quit IRC | 16:06 | |
*** michaelgugino has quit IRC | 16:17 | |
openstackgerrit | Kevin Mullery proposed openstack/watcher: Graph cluster model implementation https://review.openstack.org/362730 | 17:01 |
openstackgerrit | Digambar proposed openstack/watcher-specs: automatic triggering of action plans for AUDIT https://review.openstack.org/342473 | 17:23 |
*** diga has joined #openstack-watcher | 17:24 | |
diga | acabot: Hi, pushed new PS as per your comment - https://review.openstack.org/#/c/342473/ | 17:25 |
diga | can you see it once, lets close today only | 17:25 |
*** dtardivel has quit IRC | 17:27 | |
*** jed56 has quit IRC | 18:05 | |
*** openstackgerrit has quit IRC | 18:50 | |
*** openstackgerrit has joined #openstack-watcher | 18:50 | |
*** jimbaker has quit IRC | 18:56 | |
*** logan- has quit IRC | 18:57 | |
*** figleaf has quit IRC | 18:57 | |
*** jimbaker has joined #openstack-watcher | 18:57 | |
*** jimbaker has quit IRC | 18:57 | |
*** jimbaker has joined #openstack-watcher | 18:57 | |
*** logan- has joined #openstack-watcher | 18:57 | |
*** edleafe has joined #openstack-watcher | 18:57 | |
*** gabor_antal has quit IRC | 19:08 | |
*** gabor_antal has joined #openstack-watcher | 19:09 | |
*** openstackgerrit has quit IRC | 19:18 | |
*** thorst_ has quit IRC | 19:19 | |
*** thorst_ has joined #openstack-watcher | 19:19 | |
*** openstackgerrit has joined #openstack-watcher | 19:19 | |
*** harlowja has quit IRC | 19:32 | |
*** grumpycatt has quit IRC | 19:32 | |
*** sballe_ has quit IRC | 19:33 | |
*** aspiers has quit IRC | 19:33 | |
*** yuanying has quit IRC | 19:33 | |
*** hvprash has quit IRC | 19:33 | |
*** zigo has quit IRC | 19:33 | |
*** zigo has joined #openstack-watcher | 19:33 | |
*** zigo is now known as Guest10079 | 19:33 | |
*** aspiers has joined #openstack-watcher | 19:34 | |
*** yuanying has joined #openstack-watcher | 19:34 | |
*** sballe_ has joined #openstack-watcher | 19:44 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/watcher: Updated from global requirements https://review.openstack.org/383149 | 19:47 |
*** umbSublime_ has quit IRC | 19:48 | |
*** Daviey has quit IRC | 19:48 | |
*** umbSublime_ has joined #openstack-watcher | 19:48 | |
*** Daviey has joined #openstack-watcher | 19:48 | |
*** acabot has quit IRC | 19:56 | |
*** acabot has joined #openstack-watcher | 19:56 | |
*** openstack has joined #openstack-watcher | 20:19 | |
*** _d34dh0r53_ has joined #openstack-watcher | 20:21 | |
*** odyssey4me_ has joined #openstack-watcher | 20:21 | |
*** d34dh0r53 has quit IRC | 20:22 | |
*** odyssey4me has quit IRC | 20:22 | |
*** edleafe is now known as figleaf | 20:23 | |
*** _d34dh0r53_ is now known as d34dh0r53 | 20:24 | |
*** harlowja has joined #openstack-watcher | 20:30 | |
*** acabot has quit IRC | 20:33 | |
*** thorst_ has quit IRC | 20:33 | |
*** Daviey has quit IRC | 20:33 | |
*** umbSublime_ has quit IRC | 20:33 | |
*** sballe_ has quit IRC | 20:33 | |
*** yuanying has quit IRC | 20:33 | |
*** aspiers has quit IRC | 20:33 | |
*** Guest10079 has quit IRC | 20:33 | |
*** logan- has quit IRC | 20:33 | |
*** figleaf has quit IRC | 20:33 | |
*** gabor_antal has quit IRC | 20:33 | |
*** hvprash has joined #openstack-watcher | 20:35 | |
*** figleaf has joined #openstack-watcher | 20:38 | |
*** logan- has joined #openstack-watcher | 20:38 | |
*** Daviey has joined #openstack-watcher | 20:39 | |
*** umbSublime_ has joined #openstack-watcher | 20:39 | |
*** sballe_ has joined #openstack-watcher | 20:39 | |
*** yuanying has joined #openstack-watcher | 20:39 | |
*** aspiers has joined #openstack-watcher | 20:39 | |
*** Guest10079 has joined #openstack-watcher | 20:39 | |
*** acabot has joined #openstack-watcher | 20:39 | |
*** thorst_ has joined #openstack-watcher | 20:39 | |
*** gabor_antal has joined #openstack-watcher | 20:40 | |
*** diga has quit IRC | 20:50 | |
*** hvprash_ has quit IRC | 21:18 | |
*** thorst_ has quit IRC | 21:21 | |
*** thorst_ has joined #openstack-watcher | 21:42 | |
*** thorst_ has quit IRC | 21:46 | |
*** grumpycatt has joined #openstack-watcher | 21:52 | |
*** figleaf is now known as edleafe | 21:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!