| *** wootehfoot has joined #openstack-watcher | 00:15 | |
| *** sauvik has joined #openstack-watcher | 00:46 | |
| *** sauvik has quit IRC | 00:48 | |
| *** wootehfoot has quit IRC | 01:06 | |
| *** wootehfoot has joined #openstack-watcher | 01:08 | |
| *** vtech has quit IRC | 01:24 | |
| *** sauvik has joined #openstack-watcher | 01:35 | |
| *** ss4 has joined #openstack-watcher | 02:07 | |
| *** thorst has quit IRC | 02:13 | |
| *** thorst has joined #openstack-watcher | 02:13 | |
| *** ss4 has quit IRC | 02:17 | |
| *** thorst has quit IRC | 02:22 | |
| *** wootehfoot has quit IRC | 02:26 | |
| *** ss4 has joined #openstack-watcher | 02:28 | |
| openstackgerrit | junjie.huang proposed openstack/watcher: Workload balance migration strategy implementation https://review.openstack.org/292188 | 02:30 |
|---|---|---|
| *** wootehfoot has joined #openstack-watcher | 02:38 | |
| *** sauvik has quit IRC | 02:41 | |
| *** ss4 has quit IRC | 03:07 | |
| *** ss4 has joined #openstack-watcher | 03:19 | |
| *** thorst has joined #openstack-watcher | 03:19 | |
| *** pkoniszewski has quit IRC | 03:24 | |
| *** thorst has quit IRC | 03:27 | |
| *** aspiers_ has joined #openstack-watcher | 03:28 | |
| *** pkoniszewski has joined #openstack-watcher | 03:31 | |
| *** aspiers has quit IRC | 03:32 | |
| *** sauvik has joined #openstack-watcher | 03:41 | |
| *** sauvik has quit IRC | 03:46 | |
| *** wootehfoot has quit IRC | 04:19 | |
| *** ss4 has quit IRC | 04:28 | |
| *** sauvik has joined #openstack-watcher | 04:43 | |
| *** sauvik has quit IRC | 04:47 | |
| *** thorst has joined #openstack-watcher | 05:32 | |
| *** thorst has quit IRC | 05:37 | |
| *** sauvik has joined #openstack-watcher | 05:44 | |
| *** sauvik has quit IRC | 05:49 | |
| *** apoorv has joined #openstack-watcher | 06:17 | |
| *** vtech has joined #openstack-watcher | 06:29 | |
| *** vtech has quit IRC | 06:33 | |
| *** thorst has joined #openstack-watcher | 06:34 | |
| *** thorst has quit IRC | 06:42 | |
| apoorv | hi vincent: there's no watcher_strategies section in example of configuration file | 06:58 |
| jed56 | hi apporv | 07:00 |
| apoorv | hi | 07:00 |
| jed56 | What example of the configuration file are you looking at? | 07:01 |
| apoorv | https://factory.b-com.com/www/watcher/doc/watcher/deploy/conf-files.html | 07:03 |
| jed56 | In the watcher.conf file there is not watcher_strategies | 07:03 |
| jed56 | you can only defined the goals | 07:03 |
| jed56 | but in the current version a goal in link to a strategy | 07:04 |
| jed56 | https://github.com/openstack/watcher/blob/master/setup.cfg#L48 | 07:04 |
| jed56 | watcher_strategies is in the setup.cfg file | 07:04 |
| apoorv | so should i copy this to watcher.conf file | 07:06 |
| jed56 | no, what do you want to achieve ? :) | 07:08 |
| apoorv | i just want to create an audit template by defining a goal and for defining a goal i need a strategy entrypoint right? | 07:11 |
| jed56 | by default you can create an audit template with the DUMMY goal | 07:13 |
| jed56 | if you want to add a new goal | 07:13 |
| jed56 | you need update the section watcher_goals | 07:13 |
| jed56 | GOALS:strategy | 07:14 |
| jed56 | MY_GOAL:my_strategy_1 | 07:14 |
| jed56 | my_strategy_1 is the unique id of the strategy | 07:14 |
| jed56 | my_strategy_1 is defined in entry points in the setup.cfg file | 07:15 |
| jed56 | e.g: dummy = thirdparty.dummy:DummyStrategy | 07:15 |
| jed56 | my_strategy_1 = my_strategy_implementation_class | 07:16 |
| *** vincentfrancoise has joined #openstack-watcher | 07:32 | |
| *** thorst has joined #openstack-watcher | 07:41 | |
| *** sauvik has joined #openstack-watcher | 07:46 | |
| *** vtech has joined #openstack-watcher | 07:46 | |
| *** thorst has quit IRC | 07:47 | |
| *** brunograz has joined #openstack-watcher | 07:49 | |
| *** sauvik has quit IRC | 07:51 | |
| vtech | hello guys, vincentfrancoise, last week, we've discussed adding cold migration support to Watcher. If there still is an interest I could do it. | 07:52 |
| *** apoorv has quit IRC | 07:53 | |
| vincentfrancoise | Hi vtech | 07:54 |
| vincentfrancoise | well yes you can go for it | 07:55 |
| vincentfrancoise | just create a BP for it and ask acabot to see if you need a spec or if it can get approved as-is | 07:56 |
| vincentfrancoise | vtech: but IMHO, you won't need a full spec, the BP description should be enough | 07:57 |
| vtech | vincentfrancoise, ok, thanks. Today I will take a look on what changes will it require and submit a blueprint. | 08:06 |
| vincentfrancoise | Thanks for the initiative :) | 08:07 |
| vtech | but as far as i can see right now, I will just add elif migrattion.type=='cold' here - https://github.com/openstack/watcher/blob/78f122f2419bc3cf80cc08aca9c2f5bce5a24100/watcher/applier/actions/migration.py#L136 and add _cold_migration function. | 08:09 |
| vincentfrancoise | seems like it | 08:11 |
| vincentfrancoise | maybe a BP is overkill then | 08:11 |
| vincentfrancoise | just create a bug as whichlist then | 08:11 |
| vincentfrancoise | give me the reference | 08:12 |
| vincentfrancoise | and I'll mark it as triaged | 08:12 |
| vincentfrancoise | The only issue is that it won't be useful as long as we haven't introduce action parameters | 08:12 |
| vtech | strategy can choose the type | 08:13 |
| vtech | I can adapt ours | 08:13 |
| vincentfrancoise | oh yeah true | 08:13 |
| vincentfrancoise | I got confused with the block/non-block problem | 08:14 |
| vtech | no, block/non block only applies on live migration | 08:15 |
| vincentfrancoise | yes that's why I said I got confused :) | 08:16 |
| vtech | and i see nova helper already provides a cold migration function - https://github.com/openstack/watcher/blob/81ea37de4100070fd2ff251e94f3be5b92797e08/watcher/common/nova_helper.py#L57 so it really should be only a one file modification | 08:22 |
| vincentfrancoise | vtech: not sure but I think this function is unused and untested so you would have to add some tests for it as well | 08:24 |
| vtech | oh right, shall i still go only with a bug spec then? | 08:25 |
| vtech | I think yes | 08:27 |
| vincentfrancoise | a bug seems Ok since the action was already written in a way that was expecting the cold migration to be implemented | 08:27 |
| vincentfrancoise | you are basically filling the gap | 08:27 |
| *** apoorv has joined #openstack-watcher | 08:27 | |
| vtech | vincentfrancoise, https://bugs.launchpad.net/watcher/+bug/1564297 | 08:43 |
| openstack | Launchpad bug 1564297 in watcher "Add missing cold (non-live) VM migration support" [Undecided,New] | 08:43 |
| vincentfrancoise | vtech: done | 08:44 |
| vtech | vincentfrancoise, I don't see a way how to mark it ''wishlist' | 08:44 |
| vincentfrancoise | refresh I made a mistake at first | 08:44 |
| *** thorst has joined #openstack-watcher | 08:45 | |
| vincentfrancoise | it's in the "Importance" field | 08:45 |
| vtech | vincentfrancoise, it is only changeable by project maintainer. | 08:48 |
| vincentfrancoise | oh ok makes sense | 08:48 |
| vtech | thanks. | 08:48 |
| *** thorst has quit IRC | 08:52 | |
| apoorv | what to do after make any changes in watcher.conf | 09:09 |
| vincentfrancoise | Hi apoorv | 09:10 |
| vincentfrancoise | if you make some changes into watcher.conf | 09:10 |
| vincentfrancoise | you need to restart the Watcher services | 09:10 |
| apoorv | and also soecifying goal as DUMMY in template create works but BASIC_CONSOLIDATION doesn't | 09:13 |
| *** dtardivel has joined #openstack-watcher | 09:14 | |
| vincentfrancoise | apoorv: can you detail what doesn't actually work with BASIC_CONSOLIDATION? | 09:14 |
| apoorv | watcher audit-template-create my_three_audit BASIC_CONSOLIDATION | 09:14 |
| apoorv | it shows BASIC_CONSOLIDATION (HTTP 400) | 09:15 |
| vincentfrancoise | can you run "watcher --debug audit-template-create my_three_audit BASIC_CONSOLIDATION" and pastebin me the output please? | 09:17 |
| apoorv | http://pastebin.com/download/JcAbfQbP | 09:20 |
| vincentfrancoise | apoorv: audit template names should be unique, so can you try again with a brand new audit template name? | 09:26 |
| vincentfrancoise | apoorv: if the problem isn't solved like this, this would mean that there is a problem on the server side (it should be logged) | 09:27 |
| apoorv | where is the keyword DUMMY defined | 09:35 |
| vincentfrancoise | DUMMY is the goal we set by default in Watcher (mapped to the 'dummy' strategy) | 09:36 |
| vincentfrancoise | so unless you change the 'goals' option in the [watcher_goals] section, this is the only goal that will be available to you | 09:37 |
| vincentfrancoise | so you should have it set to something like: goals=DUMMY:dummy,BASIC_CONSOLIDATION:basic | 09:38 |
| *** sauvik has joined #openstack-watcher | 09:48 | |
| *** thorst has joined #openstack-watcher | 09:50 | |
| *** sauvik has quit IRC | 09:52 | |
| *** thorst has quit IRC | 09:56 | |
| apoorv | okay it works after restarting all watcher services | 10:20 |
| vincentfrancoise | apoorv: nice! | 10:20 |
| vincentfrancoise | apoorv: FYI, the way goals are mapped will change very soon | 10:20 |
| vincentfrancoise | see https://review.openstack.org/#/q/project:openstack/watcher+branch:master+topic:bp/get-goal-from-strategy | 10:21 |
| *** thorst has joined #openstack-watcher | 10:54 | |
| *** thorst has quit IRC | 11:01 | |
| *** Kevin_Zheng has quit IRC | 11:12 | |
| *** vtech has quit IRC | 11:37 | |
| openstackgerrit | Alexander Chadin proposed openstack/watcher-specs: Add Overload standard deviation strategy spec https://review.openstack.org/286153 | 11:41 |
| *** thorst has joined #openstack-watcher | 11:41 | |
| *** sauvik has joined #openstack-watcher | 11:49 | |
| *** alexchadin has joined #openstack-watcher | 11:49 | |
| openstackgerrit | Tomasz Kaczynski proposed openstack/watcher-specs: Add specification for Watcher Scoring Module https://review.openstack.org/289880 | 11:51 |
| *** sauvik has quit IRC | 11:54 | |
| *** apoorv has quit IRC | 12:00 | |
| *** vtech has joined #openstack-watcher | 12:04 | |
| tkaczynski | hi guys. Jenkins build failed when verifying my spec update: https://review.openstack.org/#/c/289880/ | 12:32 |
| tkaczynski | is this known issue? there is no error, just some time out | 12:32 |
| vincentfrancoise | tkaczynski: I guess this is some random failure | 12:33 |
| vincentfrancoise | you can re-run the test by issue the magic "recheck" comment | 12:33 |
| vincentfrancoise | issuing* | 12:33 |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Updated purge to now include goals and strategies https://review.openstack.org/299539 | 12:36 |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Added /strategies endpoint in Watcher API https://review.openstack.org/299899 | 12:36 |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Documentation update for get-goal-from-strategy https://review.openstack.org/299900 | 12:36 |
| *** alexchadin has quit IRC | 12:38 | |
| *** alexchadin has joined #openstack-watcher | 12:42 | |
| openstackgerrit | Vincent Françoise proposed openstack/watcher: Remove watcher_goals section from devstack plugin https://review.openstack.org/299902 | 12:43 |
| *** alexchadin has quit IRC | 12:43 | |
| *** brunograz has quit IRC | 12:44 | |
| *** brunograz has joined #openstack-watcher | 12:44 | |
| tkaczynski | vincentfrancoise sorry for a delay. how do I do this magic comment? reply with "recheck" word and some magic bot does the job? :) | 13:09 |
| vincentfrancoise | write a "recheck" comment as a reply onto your changeset's latest PS | 13:10 |
| *** sauvik has joined #openstack-watcher | 13:10 | |
| vincentfrancoise | and yeah there is a bot :) | 13:11 |
| tkaczynski | trying, we'll see :) | 13:12 |
| vincentfrancoise | It worked | 13:13 |
| vincentfrancoise | you can check if it got triggered by going to http://status.openstack.org/zuul/ | 13:14 |
| vincentfrancoise | you can then filter by openstack/watcher-specs and you'll see the job you just triggered | 13:14 |
| *** sauvik has quit IRC | 13:15 | |
| tkaczynski | bookmarked, thanks! | 13:16 |
| openstackgerrit | Vojtech Cima proposed openstack/watcher: Added cold VM migration support https://review.openstack.org/299927 | 13:21 |
| openstackgerrit | Vincent Françoise proposed openstack/python-watcherclient: Updated CLI for new goals/strategies API https://review.openstack.org/298631 | 13:26 |
| openstackgerrit | Vincent Françoise proposed openstack/python-watcherclient: Added Strategy support in Watcher CLI https://review.openstack.org/299930 | 13:26 |
| *** sauvik has joined #openstack-watcher | 13:46 | |
| *** hvprash has joined #openstack-watcher | 14:13 | |
| *** hvprash_ has joined #openstack-watcher | 14:14 | |
| *** sauvik has quit IRC | 14:15 | |
| *** sauvik has joined #openstack-watcher | 14:17 | |
| *** hvprash has quit IRC | 14:17 | |
| openstackgerrit | Vojtech Cima proposed openstack/watcher: Added cold VM migration support https://review.openstack.org/299927 | 14:44 |
| *** hvprash_ has quit IRC | 14:51 | |
| *** lrensing has joined #openstack-watcher | 14:52 | |
| vincentfrancoise | Hi vtech | 14:53 |
| vincentfrancoise | just a quick question on https://review.openstack.org/299927: did you try to run this action onto the applier to see if it works? | 14:53 |
| *** jwcroppe has quit IRC | 14:54 | |
| *** jwcroppe has joined #openstack-watcher | 14:54 | |
| vtech | vincentfrancoise, hey, you mean the underlying function? Unfortunately I didn't as i currently don't have an access to a multi node watcher deployment. | 14:56 |
| vincentfrancoise | vtech: ok so the problem with this action now is that we have no easy way to validate it | 14:58 |
| vtech | but the binding between the function and strategy should work as it is now | 14:58 |
| vincentfrancoise | are you planning on adding it to your consolidation strategy? | 14:58 |
| *** jwcroppe has quit IRC | 14:59 | |
| vtech | yes, that's the plan | 14:59 |
| *** sauvik has quit IRC | 14:59 | |
| vtech | shall I open another bug for that? | 14:59 |
| vincentfrancoise | yeah | 15:00 |
| vincentfrancoise | but link your changeset that updates the strategy to this one | 15:00 |
| vtech | I have to talk to brunograz as he might have an access to a multinode watcher deployment. | 15:00 |
| vincentfrancoise | anytime soom we might be able to add a multinode gate on review.openstack.org | 15:01 |
| vincentfrancoise | see https://review.openstack.org/#/c/299809/ | 15:01 |
| vincentfrancoise | so if you write a tempest test for your strategy, we can validate both the strategy and the cold migration: 2 birds 1 stone :) | 15:02 |
| vtech | I may just internally try to produce action list containing cold migrations to be sure that watcher_non_live_migrate_instance() works | 15:03 |
| vincentfrancoise | the advantage of doing it with a tempest test is that we can also make sure that your strategy will work at any time | 15:03 |
| vincentfrancoise | I have a meeting | 15:03 |
| vincentfrancoise | see you later ;) | 15:04 |
| vtech | I was very surprised to find out there is no function for cold migration in nova compute api. | 15:04 |
| *** vtech has quit IRC | 15:08 | |
| *** vtech has joined #openstack-watcher | 15:08 | |
| *** sauvik has joined #openstack-watcher | 15:15 | |
| *** sauvik has quit IRC | 15:30 | |
| acabot | tpeoples : hi Taylor, would you mind reviewing vincent code ? https://review.openstack.org/#/q/status:open+project:openstack/watcher+branch:master+topic:bp/get-goal-from-strategy | 15:39 |
| acabot | I would like to have it ready for mitaka | 15:40 |
| tpeoples | acabot: yes, I will review today | 15:41 |
| acabot | tpeoples : thx a lot | 15:41 |
| jed56 | vtech i think that there is a function for cold migration | 15:43 |
| jed56 | in nova api | 15:43 |
| jed56 | you have migrate() and resize() | 15:43 |
| vtech | that's in doc, but wasn't able to find it in actual code base | 15:44 |
| jed56 | in the watcher code ? | 15:45 |
| vtech | i found it in watcher code | 15:45 |
| jed56 | hmm | 15:46 |
| jed56 | https://github.com/openstack/watcher/blob/master/watcher/common/nova_helper.py#L57 | 15:46 |
| jed56 | need to be removed | 15:46 |
| jed56 | I think there was bad refactoring https://github.com/openstack/watcher/blob/77fcefbfbbfe77c0208fa47df71f4991db56da08/watcher/applier/framework/command/wrapper/nova_wrapper.py#L271 | 15:46 |
| vtech | I used this function as i expected this works | 15:46 |
| vtech | but was surprised by its implementation as i would assume, cold migration should be by default implemented in nova compute | 15:47 |
| jed56 | http://docs.openstack.org/user-guide-admin/cli_nova_migrate.html | 15:50 |
| jed56 | We should remove ef watcher_non_live_migrate_instance(self, instance_id, hypervisor_id, | 15:52 |
| jed56 | keep_original_image_name=True): | 15:52 |
| vtech | but it needs to be replaced with something else then. | 15:56 |
| vtech | let's say cold_migrate_function() | 15:57 |
| jed56 | yes you can replace more less with this code : https://github.com/openstack/watcher/blob/77fcefbfbbfe77c0208fa47df71f4991db56da08/watcher/applier/framework/command/wrapper/nova_wrapper.py#L271 ? | 15:58 |
| jed56 | what do you think ? | 15:58 |
| vtech | ok, I will check it out. I was actually expecting solution like nova.migrate() which does the whole migration without a need for "manual" volume and floating ip management. | 16:02 |
| vtech | but couldn't find migrate() function at https://github.com/openstack/nova/blob/7ba44cbb9969c1dd31c6cec26c13f98ceee7c746/nova/compute/api.py#L3350 | 16:02 |
| vtech | so then i thought your implementation of watcher_non_live_migrate_instance does exactly what i wanted, and it's implementation is so complicated because there was no other way. | 16:03 |
| jed56 | the watcher_non_live_migrate_instance was a test :) | 16:06 |
| vtech | ok just to make it clear - https://github.com/openstack/watcher/blob/77fcefbfbbfe77c0208fa47df71f4991db56da08/watcher/applier/framework/command/wrapper/nova_wrapper.py#L57 is the function which should be used for cold migration, right | 16:06 |
| vtech | ? | 16:06 |
| *** brunograz has quit IRC | 16:08 | |
| vtech | sorry, i mean this one https://github.com/openstack/watcher/blob/77fcefbfbbfe77c0208fa47df71f4991db56da08/watcher/applier/framework/command/wrapper/nova_wrapper.py#L271 | 16:08 |
| *** thorst is now known as thorst_afk | 16:08 | |
| jed56 | yes i think that the function you mention is the build in "cold migration" | 16:09 |
| vtech | I guys, I have to go now, thanks for the pointers, I will replace deprecated watcher_non_live_migrate_instance. | 16:09 |
| jed56 | thanks vtech | 16:09 |
| jed56 | good evening | 16:10 |
| vtech | you too | 16:10 |
| *** vtech has quit IRC | 16:10 | |
| jed56 | it's that you can't choose the destination | 16:11 |
| jed56 | it's was maybe the reason why i wrote watcher cold migration | 16:11 |
| *** alexchadin has joined #openstack-watcher | 16:15 | |
| alexchadin | I've changed servionica logo in presentation to english version | 16:19 |
| alexchadin | hi:) | 16:19 |
| jed56 | hi | 16:24 |
| *** wootehfoot has joined #openstack-watcher | 16:28 | |
| *** alexchadin has quit IRC | 16:32 | |
| *** alexchadin has joined #openstack-watcher | 16:33 | |
| alexchadin | offtop: is there any way to look missed messages? | 16:40 |
| vincentfrancoise | alexchadin: you mean from this channel? | 16:41 |
| alexchadin | Yes | 16:41 |
| vincentfrancoise | http://eavesdrop.openstack.org/irclogs/%23openstack-watcher/ | 16:41 |
| vincentfrancoise | everything said on this channel is logged at the address above | 16:41 |
| alexchadin | thx | 16:42 |
| *** alexchadin has quit IRC | 16:53 | |
| *** thorst_afk is now known as thorst | 16:57 | |
| *** sauvik has joined #openstack-watcher | 17:05 | |
| *** sauvik has quit IRC | 17:22 | |
| *** vincentfrancoise has quit IRC | 17:30 | |
| *** jwcroppe has joined #openstack-watcher | 17:38 | |
| *** sauvik has joined #openstack-watcher | 17:45 | |
| *** jwcroppe_ has joined #openstack-watcher | 18:11 | |
| *** jwcroppe has quit IRC | 18:12 | |
| *** vtech has joined #openstack-watcher | 18:34 | |
| *** sauvik has quit IRC | 18:50 | |
| *** esberglu has quit IRC | 18:53 | |
| *** esberglu has joined #openstack-watcher | 18:54 | |
| *** openstackgerrit has quit IRC | 19:03 | |
| *** openstackgerrit has joined #openstack-watcher | 19:04 | |
| *** dtardivel has quit IRC | 19:28 | |
| *** jwcroppe has joined #openstack-watcher | 19:49 | |
| *** jwcroppe_ has quit IRC | 19:51 | |
| *** esberglu has quit IRC | 19:57 | |
| *** sauvik has joined #openstack-watcher | 19:58 | |
| *** sauvik has quit IRC | 20:02 | |
| *** jwcroppe has quit IRC | 20:14 | |
| *** lrensing has left #openstack-watcher | 20:55 | |
| *** thorst has quit IRC | 21:07 | |
| *** thorst has joined #openstack-watcher | 21:08 | |
| *** thorst has quit IRC | 21:12 | |
| *** thorst has joined #openstack-watcher | 21:14 | |
| *** thorst has quit IRC | 21:19 | |
| *** thorst has joined #openstack-watcher | 22:16 | |
| *** thorst has quit IRC | 22:37 | |
| *** thorst has joined #openstack-watcher | 23:38 | |
| *** thorst has quit IRC | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!