*** gongysh has quit IRC | 00:13 | |
*** gongysh has joined #senlin | 00:19 | |
*** Qiming has joined #senlin | 00:46 | |
*** gongysh has quit IRC | 01:02 | |
*** zzxwill has joined #senlin | 01:42 | |
*** zzxwill has quit IRC | 01:49 | |
*** zzxwill has joined #senlin | 01:52 | |
openstackgerrit | Merged openstack/python-senlinclient: Add OSC command for senlin node-check/recover https://review.openstack.org/295114 | 01:55 |
---|---|---|
openstackgerrit | Merged openstack/python-senlinclient: Add deprecation warnings for senlin commands https://review.openstack.org/295115 | 01:55 |
*** zzxwill_ has joined #senlin | 01:56 | |
*** zzxwill has quit IRC | 01:57 | |
*** zzxwill_ has quit IRC | 02:02 | |
*** zzxwill has joined #senlin | 02:03 | |
*** jdandrea has quit IRC | 02:23 | |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Refactor osc support https://review.openstack.org/297492 | 02:33 |
*** zzxwill has quit IRC | 02:33 | |
*** zzxwill has joined #senlin | 02:41 | |
*** yanyanhu has joined #senlin | 02:54 | |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Refactor osc support https://review.openstack.org/297492 | 03:09 |
*** yuanying_ has joined #senlin | 03:35 | |
*** yuanying has quit IRC | 03:37 | |
*** yuanying has joined #senlin | 03:38 | |
*** yuanyin__ has joined #senlin | 03:40 | |
*** yuanying_ has quit IRC | 03:40 | |
*** yuanying has quit IRC | 03:43 | |
*** yuanyin__ has quit IRC | 03:49 | |
*** yuanying has joined #senlin | 03:51 | |
*** yuanying has quit IRC | 03:59 | |
*** zzxwill has quit IRC | 03:59 | |
*** zzxwill has joined #senlin | 04:00 | |
*** yuanying has joined #senlin | 04:02 | |
*** shu-mutou is now known as shu-mutou-AFK | 04:05 | |
*** yuanying has quit IRC | 04:07 | |
*** zzxwill has quit IRC | 04:14 | |
*** zzxwill has joined #senlin | 04:55 | |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Remove senlin CLI commands https://review.openstack.org/297514 | 05:10 |
Qiming | hi, we are saying goodbye to senlin command line ... | 05:10 |
Qiming | with the patch above ^ | 05:10 |
zzxwill | Why do we remove CLI? | 05:21 |
zzxwill | Sorry, I didn't get the reason from your comments of the patch. | 05:22 |
Qiming | zzxwill, http://specs.openstack.org/openstack/openstack-specs/specs/deprecate-cli.html | 05:23 |
zzxwill | Thanks for the link:) | 05:25 |
Qiming | it was just approved a few weeks ago | 05:25 |
Qiming | since this is the very first release of senlinclient | 05:26 |
Qiming | we don't want to still include the 'senlin' command line as deprecated and leave them there for two cycles for deprecation | 05:26 |
Qiming | we'd better off remove them since the very beginning | 05:27 |
zzxwill | +1 | 05:27 |
zzxwill | :) | 05:27 |
*** zzxwill has quit IRC | 05:40 | |
*** zzxwill has joined #senlin | 05:43 | |
*** zzxwill has quit IRC | 05:54 | |
*** zzxwill has joined #senlin | 06:09 | |
*** zzxwill has quit IRC | 06:27 | |
yanyanhu | hi, Qiming around? | 06:36 |
*** zzxwill has joined #senlin | 06:37 | |
yanyanhu | openstack-client is a cmd tool and it relies on the library provided by other python-**client projects, right? | 06:37 |
Qiming | yes | 06:38 |
yanyanhu | I found rally now uses clients rather than pure http API request to talk with other services | 06:38 |
yanyanhu | ok, I see | 06:39 |
Qiming | it is still a stevedore thing | 06:39 |
yanyanhu | got it | 06:39 |
Qiming | rally is doing that? | 06:39 |
Qiming | that is different from tempest | 06:39 |
yanyanhu | yes | 06:39 |
Qiming | the design philosophy of tempest, iirc, is to avoid using a per-project client | 06:39 |
Qiming | because those per-project clients may hide some bugs | 06:40 |
Qiming | that is one of the reason they have tempest-lib plugin model | 06:40 |
yanyanhu | ok, in that case, I guess the only choice is openstack-sdk | 06:40 |
Qiming | so ... seems to me, api test really should go tempest | 06:40 |
yanyanhu | since openstacklient doesn't provide lib? | 06:40 |
Qiming | but scenario tests can go either way | 06:40 |
Qiming | openstackclient goes to per-project client also | 06:41 |
yanyanhu | yes, rally is more for scenario/performance test I feel | 06:41 |
Qiming | it WAS implemented that way | 06:41 |
yanyanhu | it's not designed for API test | 06:41 |
yanyanhu | yea | 06:41 |
Qiming | there have been some efforts migrating the code to use openstacksdk | 06:41 |
Qiming | so, more accurate answer to your question would be ... openstackclient today is a mix | 06:42 |
yanyanhu | so, ideally, openstacksdk should be the lib used by rally or tempest to talk with openstack services | 06:42 |
yanyanhu | like what we are doing now in senlinclient | 06:42 |
Qiming | ideally | 06:42 |
yanyanhu | ok, I see | 06:42 |
Qiming | then the senlinclient's only purpose is to provide OSC plugins | 06:42 |
Qiming | those plugins use SDK | 06:43 |
yanyanhu | but after openstackclient switch to openstack-sdk completely, we don't need senlinclient project :) | 06:43 |
Qiming | senlinclient doesn't have to expose a senlinclient.v1.client | 06:43 |
Qiming | senlinclient's only purpose is to provide OSC plugins | 06:43 |
yanyanhu | ok, but if so, we may not be able to implement senlin plugin for rally | 06:44 |
yanyanhu | or we can use openstack-sdk | 06:44 |
Qiming | either way should work | 06:44 |
Qiming | ideally, all xxxclient should be deprecated except for the openstackclient which will talk to services via openstacksdk | 06:45 |
yanyanhu | yes | 06:45 |
yanyanhu | that's the final goal | 06:45 |
Qiming | but: 1) openstacksdk still lacks a lot of resources ... | 06:45 |
Qiming | 2) some xxxclient is very dirty ... they have too much client side logics | 06:45 |
Qiming | e.g. heatclient, doing 'get_file' parsing | 06:46 |
Qiming | e.g. novaclient, doing imageref flavorref parsing ... | 06:46 |
*** yuanying has joined #senlin | 06:46 | |
yanyanhu | ok. So now people are encouraged to use openstackclient CLI since the CLI in each individual xxxclients will be deprecated, e.g. those warnings thrower by keystone client | 06:47 |
Qiming | it was a cross-project specs approved | 06:47 |
yanyanhu | but for lib, we still don't have better choice rather than the ones provided by each xxxclient | 06:47 |
Qiming | but ... as always, there is a looooooooong way to go | 06:48 |
yanyanhu | since sdk is not mature enough | 06:48 |
yanyanhu | yes, that's true... | 06:48 |
Qiming | sdk is still introducing some disruptive changes, that concerns me a lot | 06:48 |
yanyanhu | yes | 06:48 |
Qiming | we can help 'supervise' those changes, and yell out if they don't make any senses | 06:49 |
yanyanhu | ok | 06:49 |
yanyanhu | so if we want to add rally plugin for senlin, the current choice is still python-senlinclient | 06:49 |
yanyanhu | just as you said, 'senlinclient.v1.client' | 06:50 |
Qiming | yes | 06:50 |
Qiming | that would be preferred | 06:50 |
yanyanhu | understand | 06:51 |
*** zzxwill has quit IRC | 06:55 | |
*** zzxwill has joined #senlin | 06:57 | |
*** zzxwill has quit IRC | 07:03 | |
openstackgerrit | Qiming Teng proposed openstack/senlin: Rework senlin docs based on CLI changes https://review.openstack.org/297552 | 07:11 |
*** zzxwill has joined #senlin | 07:14 | |
*** zzxwill has quit IRC | 07:24 | |
*** zzxwill has joined #senlin | 07:26 | |
*** zzxwill has quit IRC | 07:29 | |
*** zzxwill has joined #senlin | 07:31 | |
yanyanhu | hi, Qiming, just added a blueprint for senlin support in rally, https://blueprints.launchpad.net/rally/+spec/add-support-for-clustering-service-senlin | 07:39 |
yanyanhu | please help to review it, thanks | 07:39 |
Qiming | looking | 07:39 |
Qiming | so this will be committed to rally repo? | 07:40 |
yanyanhu | yes | 07:40 |
Qiming | ok | 07:40 |
yanyanhu | to make it recognize senlin service and allow user to use rally to make some basic tests for senlin | 07:41 |
yanyanhu | e.g. cluster creating/deleting | 07:41 |
Qiming | okay | 07:41 |
*** zzxwill has quit IRC | 07:55 | |
yanyanhu | hi, Qiming, about the API micro version support in Senlin, will the API requesting way change as well? | 07:59 |
Qiming | no, there will be a openstack-api-version header in any request | 07:59 |
yanyanhu | ok | 08:00 |
yanyanhu | that header is added by client? | 08:00 |
Qiming | server will check them | 08:00 |
Qiming | such a header will be used if present | 08:00 |
yanyanhu | ok, so if this header is not provided, a default version API interface will be used | 08:01 |
Qiming | yes | 08:01 |
yanyanhu | I see | 08:01 |
Qiming | let me find some references for you | 08:02 |
Qiming | https://review.openstack.org/#/c/243414/ | 08:02 |
Qiming | https://review.openstack.org/#/c/243429/ | 08:02 |
Qiming | https://review.openstack.org/#/c/243041/ | 08:02 |
yanyanhu | nice | 08:03 |
yanyanhu | thanks | 08:03 |
Qiming | still struggling on the wsgi side changes | 08:04 |
Qiming | the nova way of wsgi implementation is not the same as senlin, a lot of hacks needed | 08:04 |
Qiming | btw, please everyone help review our changes to senlinclient | 08:06 |
Qiming | just talked to ttx about rebranching senlinclient | 08:06 |
Qiming | he is not very supportive, for good reasons, because they are the guardians of library releases | 08:06 |
yanyanhu | ok | 08:07 |
Qiming | I'm gonna write an email to him and doug to explain the necessity to rebase senlinclient stable release | 08:07 |
yanyanhu | maybe we postpone it to next cycle? | 08:07 |
Qiming | that means we will have to keep 'deprecated' warning for at least two cycles | 08:07 |
yanyanhu | yes, that's terrible... | 08:07 |
Qiming | it will bring us more troubles than any conveniences | 08:08 |
Qiming | since we already have OSC fully supported | 08:08 |
Qiming | let's strive to drop our own shell asap | 08:08 |
yanyanhu | ok | 08:08 |
Qiming | my heart is bleeding ... when deleting those modules ... | 08:09 |
yanyanhu | Understand, sorry for your lost. But it's good for people :P | 08:11 |
*** zzxwill has joined #senlin | 08:13 | |
yanyanhu | just like words in movie... "this sacrifice is for entire human being" | 08:14 |
* Qiming is feeling speechless | 08:17 | |
*** zzxwill has quit IRC | 08:27 | |
openstackgerrit | Merged openstack/python-senlinclient: Refactor osc support https://review.openstack.org/297492 | 08:30 |
openstackgerrit | Merged openstack/python-senlinclient: Remove senlin CLI commands https://review.openstack.org/297514 | 08:30 |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Rename cluster scaling command https://review.openstack.org/297572 | 08:35 |
*** zzxwill has joined #senlin | 08:35 | |
Qiming | yanyanhu, dixiaoli, please help review this ^ | 08:35 |
Qiming | the command 'openstack cluster scale in --count 2' is weird | 08:36 |
Qiming | I'm proposing to rename it as 'openstack cluster shrink --count 2' | 08:36 |
yanyanhu | directly using 'openstack cluster scale-in --count 2'? | 08:37 |
Qiming | no, openstackclient will substitute any '-' or '_' in the command | 08:37 |
yanyanhu | oh | 08:37 |
Qiming | I tried 'openstack --help', there is no such a command containing '_' or '-' | 08:38 |
yanyanhu | or maybe just scalein | 08:38 |
Qiming | so having an 'in' in a command is very annoying, it can be interpreted as a shell reserved word, for example | 08:38 |
yanyanhu | oh, the cli of senlinclient will be deprecated | 08:39 |
yanyanhu | so user will not be aware of it in future | 08:39 |
yanyanhu | forgot this | 08:39 |
yanyanhu | shrink is ok for me | 08:39 |
Qiming | neither 'scalein' nor 'scaleout' is an English word | 08:39 |
Qiming | I'd like scale-in better, if it is supported | 08:40 |
yanyanhu | shrink looks look. but I'm not sure whether there is better choice than grow :) | 08:42 |
yanyanhu | maybe we can get suggestion from cschulz :) | 08:43 |
Qiming | tried search google and bing | 08:44 |
Qiming | grow seems the best choice to my knowledge | 08:44 |
Qiming | I also considered 'expand', 'increase', 'inflate' (vs. 'deflate') ... | 08:44 |
yanyanhu | how about enlarge? | 08:46 |
yanyanhu | but it is similar to inflate? | 08:46 |
Qiming | yep | 08:46 |
Qiming | let me try some other dictionaries | 08:52 |
Qiming | maybe expand is better | 08:56 |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Rename cluster scaling command https://review.openstack.org/297572 | 08:58 |
Qiming | enlarge is a transitive verb, rarely mentioned in any dicts | 08:59 |
*** zzxwill has quit IRC | 09:00 | |
yanyanhu | ok, actually I just used youdao dict which is from net :) | 09:00 |
yanyanhu | not that professinal | 09:00 |
yanyanhu | s/net/netyi | 09:01 |
Qiming | it depends on context | 09:02 |
yanyanhu | hi Qiming , the commit message need to be updated as well | 09:02 |
Qiming | okay | 09:02 |
*** zzxwill has joined #senlin | 09:02 | |
openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Rename cluster scaling command https://review.openstack.org/297572 | 09:02 |
*** openstackstatus has joined #senlin | 09:09 | |
*** ChanServ sets mode: +v openstackstatus | 09:09 | |
*** zzxwill has quit IRC | 09:17 | |
*** zzxwill has joined #senlin | 09:20 | |
yanyanhu | will leave a little earlier today, ttyl | 09:23 |
Qiming | bb | 09:23 |
*** yanyanhu has quit IRC | 09:23 | |
*** lixinhui_ has joined #senlin | 09:24 | |
openstackgerrit | Merged openstack/python-senlinclient: Rename cluster scaling command https://review.openstack.org/297572 | 09:26 |
*** zzxwill has quit IRC | 09:27 | |
*** zzxwill has joined #senlin | 09:39 | |
*** zzxwill has quit IRC | 09:39 | |
openstackgerrit | Qiming Teng proposed openstack/senlin: Rework senlin docs based on CLI changes https://review.openstack.org/297552 | 09:43 |
*** ddeja has quit IRC | 09:44 | |
*** ddeja has joined #senlin | 09:44 | |
*** Qiming has quit IRC | 10:03 | |
*** Qiming has joined #senlin | 11:02 | |
*** lixinhui_ has quit IRC | 11:47 | |
openstackgerrit | Qiming Teng proposed openstack/senlin: Add versioned method support https://review.openstack.org/297639 | 11:48 |
*** lixinhui_ has joined #senlin | 13:29 | |
*** lixinhui_ has quit IRC | 14:08 | |
*** jdandrea has joined #senlin | 15:41 | |
*** Qiming has quit IRC | 16:20 | |
*** dhellmann has quit IRC | 23:24 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!