Friday, 2016-07-29

« Thursday, 2016-07-28 Index Saturday, 2016-07-30 »
*** singhj has quit IRC00:08
*** jamielennox is now known as jamielennox|away00:49
*** zhurong has joined #openstack-sdks01:12
*** yanyanhu has joined #openstack-sdks01:15
*** bvandewa has quit IRC01:28
*** jamielennox|away is now known as jamielennox01:34
*** hoangcx has joined #openstack-sdks02:04
openstackgerritQiangTang proposed openstack/osc-lib: Prevent null key setting for property  https://review.openstack.org/34530502:05
*** zhurong has quit IRC02:24
*** zhurong has joined #openstack-sdks02:24
*** jamielennox is now known as jamielennox|away02:32
*** gouthamr_ has quit IRC02:51
openstackgerritQiangTang proposed openstack/osc-lib: Add reno for osc-lib release notes management  https://review.openstack.org/34821002:53
*** yuanying_ has quit IRC02:57
*** mbonell has joined #openstack-sdks02:58
*** thrash is now known as thrash|g0ne03:00
openstackgerritQiangTang proposed openstack/osc-lib: Add reno for osc-lib release notes management  https://review.openstack.org/34821003:06
openstackgerritQiangTang proposed openstack/python-openstackclient: Update the description of project in releasenotes.  https://review.openstack.org/34870403:13
openstackgerritMerged openstack/osc-lib: Updated from global requirements  https://review.openstack.org/34865503:34
*** yuanying has joined #openstack-sdks03:51
*** hieulq has joined #openstack-sdks03:51
*** mbonell has quit IRC04:06
*** NanKe has joined #openstack-sdks04:59
openstackgerritQiangTang proposed openstack/osc-lib: Add reno for osc-lib release notes management  https://review.openstack.org/34821005:01
openstackgerritMerged openstack/keystoneauth: Updated from global requirements  https://review.openstack.org/34862905:09
*** amitkqed has quit IRC05:19
*** yanyanhu has quit IRC05:19
*** amitkqed has joined #openstack-sdks05:20
*** yanyanhu has joined #openstack-sdks05:21
*** amotoki has joined #openstack-sdks05:46
openstackgerritMerged openstack/python-openstackclient: Updated from global requirements  https://review.openstack.org/34868005:51
*** gouthamr has joined #openstack-sdks06:01
*** gouthamr has quit IRC06:10
*** e0ne has joined #openstack-sdks06:12
*** hoangcx has quit IRC06:42
*** bvandewa has joined #openstack-sdks07:16
*** yanyanhu has quit IRC07:17
*** gildub has quit IRC07:19
*** sheel has joined #openstack-sdks07:32
*** fzdarsky|afk has joined #openstack-sdks07:44
*** bvandewa has quit IRC07:45
*** e0ne has quit IRC08:00
*** fzdarsky has joined #openstack-sdks08:11
*** fzdarsky|afk has quit IRC08:11
*** bvandewa has joined #openstack-sdks08:32
*** NanKe has quit IRC08:42
*** NanKe has joined #openstack-sdks08:44
*** hoangcx has joined #openstack-sdks08:46
*** yanyanhu has joined #openstack-sdks08:51
*** e0ne has joined #openstack-sdks08:52
*** amotoki has quit IRC08:56
*** amotoki has joined #openstack-sdks09:00
*** e0ne has quit IRC09:05
*** yanyanhu has quit IRC09:20
*** bvandewa has quit IRC09:27
*** cdent has joined #openstack-sdks09:49
*** Klumben has quit IRC09:54
*** Romanenko_K has joined #openstack-sdks09:56
*** Klumben has joined #openstack-sdks09:58
*** e0ne has joined #openstack-sdks10:01
*** zhurong has quit IRC10:03
*** Qiming has quit IRC10:08
*** e0ne has quit IRC10:12
*** Qiming has joined #openstack-sdks10:13
*** sheel has quit IRC10:16
*** gildub has joined #openstack-sdks10:16
*** sdague has joined #openstack-sdks10:19
*** Qiming has quit IRC10:25
*** Qiming has joined #openstack-sdks10:27
*** hoangcx has quit IRC10:33
*** Qiming has quit IRC10:46
*** Qiming has joined #openstack-sdks10:52
*** Qiming has quit IRC10:57
*** e0ne has joined #openstack-sdks10:58
*** e0ne has quit IRC11:01
*** Qiming has joined #openstack-sdks11:01
*** Romanenko_K has quit IRC11:10
*** lucasagomes is now known as lucas-hungry11:14
*** rtheis has joined #openstack-sdks11:21
*** Romanenko_K has joined #openstack-sdks11:24
*** tangchen has quit IRC11:40
*** tangchen_ has quit IRC11:40
*** thrash|g0ne is now known as thrash11:41
*** tangchen_ has joined #openstack-sdks11:42
*** tangchen has joined #openstack-sdks11:42
*** e0ne has joined #openstack-sdks11:51
*** tangchen has quit IRC11:58
*** tangchen_ has quit IRC11:58
*** lucas-hungry is now known as lucasagomes12:20
*** edleafe is now known as figleaf12:21
*** NanKe has quit IRC12:41
*** NanKe has joined #openstack-sdks12:46
*** _elmiko is now known as elmiko12:48
*** edmondsw has joined #openstack-sdks12:54
*** NanKe has quit IRC13:05
openstackgerritMerged openstack/python-openstackclient: Update the description of project in releasenotes.  https://review.openstack.org/34870413:13
*** gouthamr has joined #openstack-sdks13:18
openstackgerritMerged openstack/python-openstackclient: Pass security group id to novaclient while adding security group to server  https://review.openstack.org/34846313:21
*** cleong has joined #openstack-sdks13:22
*** michaelgugino has joined #openstack-sdks13:26
*** cdent_ has joined #openstack-sdks13:33
*** cdent has quit IRC13:36
*** cdent_ is now known as cdent13:36
*** gildub has quit IRC13:53
*** edmondsw has quit IRC13:56
*** amotoki has quit IRC13:58
openstackgerritDean Troyer proposed openstack/python-openstackclient: osc-lib: shell  https://review.openstack.org/34611313:59
*** edmondsw has joined #openstack-sdks14:34
*** cdent has quit IRC14:36
*** zhurong has joined #openstack-sdks14:40
*** david-lyle has joined #openstack-sdks14:42
*** zhurong_ has joined #openstack-sdks14:44
*** zhurong has quit IRC14:46
*** david-lyle has quit IRC14:47
*** cdent has joined #openstack-sdks14:49
*** mbonell has joined #openstack-sdks15:03
*** mbonell has quit IRC15:09
*** singhj has joined #openstack-sdks15:10
*** e0ne has quit IRC15:14
*** mbonell has joined #openstack-sdks15:17
*** david-lyle has joined #openstack-sdks15:23
*** zhurong_ has quit IRC15:28
openstackgerritMerged openstack/python-openstackclient: Implement network rbac create and delete commands  https://review.openstack.org/34670515:47
*** sdague has quit IRC15:59
*** mbonell has quit IRC16:06
*** mbonell has joined #openstack-sdks16:06
*** bvandewa has joined #openstack-sdks16:14
*** bvandewa has quit IRC16:15
*** mbonell has quit IRC16:24
*** mbonell has joined #openstack-sdks16:24
openstackgerritMerged openstack/osc-lib: Add reno for osc-lib release notes management  https://review.openstack.org/34821016:24
*** jamielennox|away is now known as jamielennox16:32
*** mbonell has quit IRC16:39
*** lucasagomes is now known as lucas-dinner16:41
*** e0ne has joined #openstack-sdks16:45
*** michaelgugino has quit IRC16:45
*** mbonell has joined #openstack-sdks16:45
*** e0ne has quit IRC16:46
openstackgerrithenry-nash proposed openstack/python-openstackclient: Add support for domain specific roles  https://review.openstack.org/28355416:50
*** mbonell has quit IRC16:50
*** mbonell has joined #openstack-sdks16:51
*** thrash is now known as thrash|biab16:56
*** tangchen has joined #openstack-sdks16:56
*** tangchen_ has joined #openstack-sdks16:56
*** tangchen_ has quit IRC17:01
*** tangchen has quit IRC17:01
*** mbonell has quit IRC17:09
*** sdague has joined #openstack-sdks17:17
openstackgerritBrian Haley proposed openstack/python-openstackclient: Add Subnet service-types to subnets  https://review.openstack.org/34297617:19
*** mbonell has joined #openstack-sdks17:21
*** singhj has quit IRC17:23
*** jamielennox is now known as jamielennox|away17:35
*** singhj has joined #openstack-sdks17:38
*** Romanenko_K has quit IRC17:41
*** singhj has quit IRC17:42
*** mbonell has quit IRC17:49
*** e0ne has joined #openstack-sdks17:50
*** mbonell has joined #openstack-sdks17:57
*** singhj has joined #openstack-sdks18:08
*** singhj has quit IRC18:13
openstackgerritMerged openstack/js-openstack-lib: Spec for ECMAScript 2015 and Transpiler support  https://review.openstack.org/33245818:21
openstackgerritPolly Zhou proposed openstack/microversion-parse: Convert dict headers and Webob headers into lowercase.  https://review.openstack.org/34799318:23
*** singhj has joined #openstack-sdks18:26
*** thrash|biab is now known as thrash18:34
rtheisdtroyer: ping18:49
*** mbonell has quit IRC18:50
*** mbonell has joined #openstack-sdks18:50
dtroyerrtheis: hey18:53
rtheisdtroyer: neutron has several apis to list resources for a given resource...18:54
rtheisdtroyer: something like "router list port"18:54
rtheisdtroyer: but it appears that "router list" would hide a "router list port" command18:55
dtroyerI'd look at "router list —port" for that18:55
dtroyeror "port list —router <router>"18:55
dtroyerMy thinking is along the lins of "what is it you are listing?"18:55
*** mbonell has quit IRC18:55
dtroyerthat is where the command should be, with filtering18:56
dtroyerso my second example, not the first18:56
dtroyerit is really the nested resource problem that we have not solved yet18:56
rtheisdtroyer: okay, thanks18:57
rtheisdtroyer: so the nested resource problem is what hides the "router list port" command?18:57
*** mctaylor has joined #openstack-sdks18:58
rtheisThe command hiding is what is prompting a new "enumerate" action being considered in https://review.openstack.org/#/c/340624/1818:59
dtroyerI sw that, and don't like it19:00
dtroyerI'm not talking about command hiding, which I assume is in regards to help and parsing?19:00
dtroyerI am talking about the general problem of a resource A that has a one-to-many relationship with another resource B19:00
dtroyerA==router, B==port in your example19:00
rtheisokay19:01
dtroyerhow to cleanly list all of the Bs associated with a specific A19:01
dtroyerthere is a problem of output format, particualrly when you want output to contain multiple A resources19:01
dtroyerbut the general rule of the command format to do the listing19:01
dtroyerthat's why my first choice is to start with what it is you are actually listing (B, or ports) and aply A as a filter19:02
rtheisokay, thanks19:02
dtroyerwe have this problem in a number of places19:03
rtheisIt seems like in this case "subports" cannot be listed on their own without the owning "network trunk".19:04
dtroyerright now, that's where we put a list of B's into a single field in list and show outputs19:04
dtroyerso subport list —trunk <trunk> would be required?19:04
rtheisI think so19:04
rtheismaybe turn the option into positional argument?19:05
dtroyeris a subport a regular port that has been assigned to a trunk?  oris it differnt?19:05
*** cleong has quit IRC19:05
dtroyerpositionals are used for the resource names in the command19:05
rtheisa subport is a regular port with additional metadata wrapping it19:06
rtheisthat is stored with the trunk19:07
dtroyerok, so should we always call it 'port' then?19:07
*** fifieldt has quit IRC19:07
dtroyeror is subport going to be what users will expect?19:07
rtheistrunks have a parent port too19:07
rtheisso they may expect subport19:07
dtroyerok.  so then would it make sense to have 'subport list —trunk <trunk>' as a distinction?19:08
dtroyerI'm not sure how I feel about that19:08
rtheisyeah19:09
dtroyerbut if ports are called subports everywhere they are associated with a trunk (ignoring parent ports) then it might make sense to users19:09
rtheisthese subports would still show up with 'port list'19:10
rtheisbut no subport specific information for them19:11
dtroyerso then a —trunk filter on port list would be more consistent with the other list commands19:11
dtroyerthat would get the subport info19:11
*** bnemec is now known as beekneemech19:12
rtheisyes that's another option19:12
rtheisbut then we bridge the trunk support between osc and neutron osc plugin19:12
dtroyerso that hurries up our work to getting plugins to modify built-in commands, which has a lot of scary implications to it given we are dealing with credentials and trusting plugins19:13
rtheisright19:13
rtheisin the meantime 'subport list' could make sense too by itself19:13
rtheisit could list all subports across all trunks19:14
dtroyerright, and would be easy enough to transplant if/when it moves into the osc repo19:14
dtroyerinto the port command19:14
rtheissure19:14
dtroyerI could imagine where you might want a subport list filtered by all trunks on a specific router.19:15
rtheissure19:15
dtroyerbut that can still fit into port list19:15
rtheisdtroyer: are you okay with 'subport' as the object or would you prefer 'network subport' ?19:17
dtroyerI really want to minimize the use of the API names like that… we have port already, subport seems the right match19:18
dtroyerI can't think of where else port or subport would be used…19:18
*** fifieldt has joined #openstack-sdks19:18
dtroyerfamous last words...19:18
rtheis:)19:18
rtheissounds good19:18
rtheisthanks for working through this with me19:19
*** fzdarsky is now known as fzdarsky|afk19:19
*** fzdarsky|afk has quit IRC19:19
dtroyernp, the networking stuff gets gnarly, and helps sanity check what we already have19:19
rtheisdefinitely19:19
*** bvandewa has joined #openstack-sdks19:22
*** edmondsw has quit IRC19:35
*** bvandewa has quit IRC19:35
*** e0ne has quit IRC20:04
*** bvandewa has joined #openstack-sdks20:12
rtheisdtroyer: still around?20:26
dtroyerrtheis: yup20:26
rtheismore trunk questions20:27
rtheishttps://review.openstack.org/#/c/340624/1820:27
rtheissee my latest comment20:27
rtheison add/remove versus set/unset20:27
rtheisI'm not sure which is best way to go20:28
dtroyerfrankly, if trunk is really metadata on ports, it should have been implemented like that.20:30
dtroyerin this case, add/remove fits the model better20:31
dtroyerbut we didn't anticipate adding multiple in one command20:31
dtroyerport set —trunk <trunk> <port> seems cleaner to me20:32
dtroyerstill doesn't get multiple in one command though20:32
dtroyerhow important is that?20:32
rtheisI'm not sure20:32
rtheistrunk create --subport20:33
rtheissubport can be set on create20:33
dtroyerso how about this?20:33
rtheiswhich is why I was thinking trunk set --subport20:33
dtroyertrunk add subport <trunk> <subport> <subport> …20:33
dtroyermultiple positionals?20:33
dtroyermin == 220:33
dtroyerwe havent done that anywhere else yet, but it fits the model20:34
rtheisThe problem is that each subport needs segmentation ID and type set20:34
dtroyerwould extend the rules slightly20:34
dtroyer(no doc in that review yet)  how is that handled in create?20:35
dtroyermultiple options?20:35
rtheisone option20:35
dtroyercan we use the same syntax to make it one positional then?20:35
rtheis--sub-port <port=,segmentation-type=,segmentation-id=>20:35
dtroyerah, but there is an additional port name/id?20:36
rtheisyes port= is name/id20:36
rtheisThat option can be repeated20:36
*** cardeois_ has joined #openstack-sdks20:36
dtroyerso since the add is more than just an add, it is an add + set, it makes this hard ;)20:36
dtroyerthere is a convention for doing this, but I'm not sure if argparse can handle it20:37
dtroyertrunk add —opt1 aaa —opt 2 bbb port1 —opt1 ccc —opt2 ddd port220:38
dtroyerso it is processed in command order, and whatever options are in place get used when the positional comes20:38
dtroyerI think tar does this20:38
dtroyerit's obscure though and would be totally confusing for many users20:38
rtheisI agree20:39
dtroyerright now I don't see a clean way to do more than one add per command20:39
rtheisI didn't understand your comment about add + set20:39
dtroyerthe real operation being performed here is adding a subport to a trunk, and setting some metadata on that combination.20:40
rtheisah yes, I understand now20:40
dtroyerin pure port terms, it is setting a trunk and other metadata on a port20:40
dtroyerwhich I still think is cleaner20:40
dtroyerbut get a lot of pushback from folks on… volume-type had this a while back20:41
dtroyerthey wanted a new resource name for what amounted to more properties on volume-type, because it was a separate table in the database20:41
*** gouthamr has quit IRC20:41
dtroyerthat stuff shouldn't leak20:41
rtheisdtroyer: so with --sub-port <port=,segmentation-type=,segmentation-id=> on 'network trunk create' ... could they do the same on 'network trunk set'20:43
rtheisor is that not good either?20:43
dtroyercreate and set should usually be very similar, so that should be ok.  is create already like that?20:43
rtheisyes20:44
rtheisyou can do 'network trunk create --sub-port <port=,segmentation-type=,segmentation-id=>'20:44
rtheisand repeat --sub-port20:44
dtroyeroh, well then… even if add has the options, set needs it too. and that then makes adding it to add unnecessary ;)20:45
dtroyercreate is nearly always a subset of set20:45
dtroyerin theory anyway20:45
rtheissure20:45
rtheisthanks dtroyer20:46
dtroyeranytime20:46
*** cdent has quit IRC20:50
*** cardeois_ is now known as cardeois21:00
*** cdent has joined #openstack-sdks21:04
*** bvandewa has quit IRC21:04
*** singhj has quit IRC21:07
*** singhj has joined #openstack-sdks21:11
*** cdent has quit IRC21:12
*** singhj has quit IRC21:23
*** singhj has joined #openstack-sdks21:25
*** elmiko is now known as _elmiko21:32
*** bvandewa has joined #openstack-sdks21:35
openstackgerritDean Troyer proposed openstack/python-openstackclient: Add shell integration test  https://review.openstack.org/34907921:40
*** dims_ has quit IRC21:44
*** dims has joined #openstack-sdks21:46
*** singhj has quit IRC21:48
*** rtheis has quit IRC21:49
*** singhj has joined #openstack-sdks21:50
*** singhj has quit IRC21:56
*** bvandewa has quit IRC21:57
*** singhj has joined #openstack-sdks21:58
*** singhj has quit IRC22:03
*** mctaylor has quit IRC22:14
*** singhj has joined #openstack-sdks22:15
*** cardeois has quit IRC22:34
*** mctaylor has joined #openstack-sdks22:43
*** mctaylor has quit IRC22:44
*** mctaylor has joined #openstack-sdks22:45
*** gouthamr has joined #openstack-sdks22:45
*** singhj has quit IRC22:50
*** singhj has joined #openstack-sdks22:52
*** GreenBlood has quit IRC22:56
*** kuzko has quit IRC22:56
*** sdague has quit IRC22:59
openstackgerritMerged openstack/keystoneauth: Fix arguments to _auth_required()  https://review.openstack.org/34852322:59
*** mctaylor has quit IRC23:08
*** singhj has quit IRC23:08
*** thrash is now known as thrash|wknd23:32
« Thursday, 2016-07-28 Index Saturday, 2016-07-30 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!