| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-sdks - 2017-03-28 | |||
| 08:38:59 | reedip | If they have platinum , then .... you can have better number of resources in default | |
| 08:40:15 | openstackgerrit | Reedip proposed openstack/python-openstackclient master: Add "qos-policy" option to "port create" & "port set" https://review.openstack.org/419236 | |
| 08:40:28 | reedip | ralonsoh : you forgot to add yourself as the author ^^ | |
| 08:40:40 | ralonsoh | No, no problem | |
| 08:41:07 | ralonsoh | reedip: it was just a forgotten patch I wanted to be fixed | |
| 08:41:13 | reedip | I added you as well ralonsoh ... you are working on it right now | |
| 08:41:29 | reedip | yeah, you are working on it.... so I added you :) | |
| 08:41:46 | ralonsoh | reedip: okidoki!! | |
| 08:42:04 | rabel | reedip: in this case the customer would have one openstack-project, not? so you could just adjust the quotas for the project. | |
| 08:42:16 | rabel | i still don't understand. sorry ;) | |
| 08:42:27 | reedip | rabel : see it from a Public cloud perspective | |
| 08:42:37 | reedip | where multiple users can request for quotas | |
| 08:42:43 | reedip | something like AWS ? | |
| 08:43:04 | rabel | never worked with aws. | |
| 08:43:32 | rabel | but quotas in openstack are always project-based, right? | |
| 08:43:42 | reedip | rabel : neither did I , but I checked their business model | |
| 08:43:47 | reedip | rabel : yes | |
| 08:44:24 | rabel | reedip: so one project has one set of quotas. | |
| 08:44:29 | reedip | rabel : yes | |
| 08:44:54 | rabel | reedip: and additionally you have one cloud-wide default set of quotas, which is applied when a new project is created. | |
| 08:45:40 | rabel | reedip: or wait | |
| 08:45:55 | reedip | rabel : thats the thing ... you can also have a default quota for a tenant | |
| 08:46:43 | rabel | reedip: and when this is used? when i create a new tenant with --parent to that other tenant? | |
| 08:50:45 | RuiChen | dtroyer: stevemar hi, I'm drafting new commands to add network interface for a server, I have two object name "interface" and "nic" now, which one you favor? the command looks like: server add interface <server> [--network | --port] | |
| 09:00:11 | rabel | reedip: and also in this context: i do not understand, what a quota class is :D | |
| 09:00:39 | reedip | sorry , in another meeting rabel :) | |
| 09:00:46 | reedip | will be back in sometime | |
| 09:00:53 | rabel | reedip: np | |
| 09:00:57 | rabel | reedip: and thank you! | |
| 09:18:53 | reedip | rabel : hi | |
| 09:19:29 | rabel | reedip: welcome back :) | |
| 09:19:36 | reedip | currently there is no discrimination between tenants for default quotas in vanilla flavor of Openstack | |
| 09:19:52 | reedip | but it can be done..... | |
| 09:19:57 | reedip | quota class in SDK ? | |
| 09:22:16 | rabel | yes, i think so | |
| 09:22:35 | rabel | reedip: i have seen it in the openstackclient | |
| 09:22:44 | rabel | p.e. openstack quota set --class | |
| 09:23:00 | reedip | ohh that ? its similar to the project attribute | |
| 09:23:17 | reedip | but I am also confused a bit about its implementation, IMHO | |
| 09:24:03 | rabel | what i already found out: you can use this to set default quotas, by using --class default | |
| 09:24:33 | reedip | is it ? :D | |
| 09:24:44 | rabel | yes :D | |
| 09:25:57 | rabel | while "openstack quota set default" without --class does not work. i think it probably tries to set quotas for a project called "default". but it doesn't fail either, it just does not have the desired outcome ;) | |
| 09:27:44 | reedip | rabel whoa... wait ! | |
| 09:28:51 | reedip | rabel : okay, I see your point | |
| 09:29:08 | reedip | So default is type of a Class | |
| 09:29:15 | reedip | and we can create more classes | |
| 09:29:35 | reedip | and set quotas as per classes on the tenants (??) | |
| 09:30:37 | rabel | that would be reasonable, but i think the last point is not right. because you can only either specify a project or a class (at least with openstackclient) | |
| 09:31:10 | rabel | actually i get the feeling, that the whole quota concept has some weaknesses ;) | |
| 09:35:58 | reedip | rabel thats y I had the ?? in my understanding | |
| 09:37:39 | rabel | reedip: do you know who could clarify this? | |
| 09:38:03 | reedip | rabel : as of now, I am not sure myself.. I am investigating your point | |
| 09:38:37 | rabel | reedip: k, thank you | |
| 09:39:40 | reedip | rabel : well , the openstacksdk states that network/v2/quota.py QuotaDefault should be picked up when setting the default quota | |
| 09:40:48 | reedip | rabel : interesting point is that --class only stores True or False :D | |
| 09:41:18 | reedip | rabel : python-openstackclient/openstackclient/common/quota.py | |
| 09:41:27 | reedip | Line#276 | |
| 09:42:43 | reedip | and for network , it should fail ..... ??? !!! | |
| 09:42:51 | rabel | reedip: yes, the class is the positional argument instead of project in this case | |
| 09:43:09 | reedip | Line #330 states that it ignores the Quota class for networks :D | |
| 09:43:22 | rabel | reedip: you're right | |
| 09:43:25 | rabel | :D | |
| 09:44:56 | reedip | rabel : so ... we need to find how the quota_classes for compute client and volume client are to be done | |
| 09:45:55 | reedip | but cant do anything about network :) | |
| 09:47:02 | reedip | leaving ... catch u in 2 hours , I guess | |
| 09:47:37 | rabel | kk, cu | |
| 09:47:46 | rabel | i'll have a look at nova | |
| 11:46:07 | rabel | someone could have a loot at https://review.openstack.org/#/c/444924/ ? it is without review for a week now | |
| 12:14:26 | reedip | o/ | |
| 12:24:32 | openstackgerrit | Monty Taylor proposed openstack/os-client-config master: Stop special-casing idenity catalog lookups https://review.openstack.org/450259 | |
| 13:32:15 | kuzko_ | hello, I encounter an issue using the openstack python sdk giving me an OpenSSL.SSL.Error: [('SSL routines', 'SSL3_READ_BYTES', 'sslv3 alert handshake failure')] what does that imples? | |
| 14:26:38 | openstackgerrit | Akihiro Motoki proposed openstack/python-openstacksdk master: Filtering support by is_router_external to network resource https://review.openstack.org/450152 | |
| 14:59:03 | openstackgerrit | Akihiro Motoki proposed openstack/python-openstacksdk master: Filtering support by is_router_external to network resource https://review.openstack.org/450152 | |
| 15:13:31 | dtroyer | stevemar, ankur-gupta-f1, RuiChen: could you guys give https://review.openstack.org/450452 and https://review.openstack.org/450453 a look? They handle the problem with the current master os-client-config that needs to be released in both osc and osc-lib before we can release another o-c-c. | |
| 15:13:44 | dtroyer | that fixes our functional-tips job | |
| 15:15:34 | ankur-gupta-f4 | K will check in a bit after next meeting | |
| 15:19:40 | sfinucan | dtroyer: RE [1], do you want me to squash all those fixup patches? [1] https://review.openstack.org/#/q/topic:sphinx15+project:openstack/python-openstackclient | |
| 15:22:54 | dtroyer | sfinucan: if you have to touch any of them again I would, otherwise I'm on the fence. I know I prefer to review stuff like that (all-docs, etc) together because it is just simpler for me mentally. | |
| 15:23:30 | sfinucan | dtroyer: I have one or two to clean up so yeah, I'll squash them | |
| 15:23:35 | dtroyer | We normally don't need to be concerned about the resource usage of running all of those duplicate jobs for no code changes | |
| 15:23:46 | dtroyer | except in deadline weeks | |
| 15:23:54 | dtroyer | but it is good to be aware of that | |
| 15:24:14 | dtroyer | ok, thanks | |
| 15:24:32 | sfinucan | dtroyer: Yeah, I was thinking along the lines of keeping the commits self-contained | |
| 15:24:46 | ankur-gupta-f4 | sfinucan: o/ | |
| 15:24:47 | sfinucan | ...but that doesn't really make sense when I need all of them to switch the warning-is-error toggle | |
| 15:24:59 | sfinucan | ankur-gupta-f4: o/ | |
| 15:25:04 | dtroyer | right. that's always a balance, but with docs I push that line waaaaay over the other way most of the time. | |
| 15:25:19 | sfinucan | ankur-gupta-f4: what happened to ankur-gupta-{f1,f2,f3}? :) | |
| 15:25:34 | dtroyer | right. you can argue that they are all realted to that and I'd buy it :) | |
| 15:25:43 | sfinucan | job | |
| 15:26:36 | ankur-gupta-f4 | I believe i have IRC clients running on various machines. Probably | |
| 15:32:32 | reedip | he is a Drone ! | |
| 15:32:50 | reedip | he is Ankur in a Shell :D | |
| 15:54:28 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Remove local fork of apidoc https://review.openstack.org/447589 | |
| 15:54:29 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: Use Sphinx 1.5 warning-is-error https://review.openstack.org/446629 | |
| 15:54:29 | openstackgerrit | Stephen Finucane proposed openstack/python-openstackclient master: doc: Correct Sphinx warnings https://review.openstack.org/447590 | |
| 15:55:52 | sfinucan | dtroyer: ^ | |
| 15:56:09 | sfinucan | reedip: Shh! No spoilers! | |