*** jerrygb has quit IRC | 00:17 | |
*** muralia has joined #openstack-containers | 00:50 | |
*** muralia has quit IRC | 00:54 | |
*** tonanhngo has joined #openstack-containers | 01:03 | |
*** tonanhngo has quit IRC | 01:08 | |
*** zz_dimtruck is now known as dimtruck | 01:12 | |
*** sergmelikyan has joined #openstack-containers | 01:30 | |
*** sergmelikyan has quit IRC | 01:34 | |
*** hieulq has joined #openstack-containers | 01:52 | |
*** anushkrishnamurt has joined #openstack-containers | 02:00 | |
*** kevin has joined #openstack-containers | 02:34 | |
*** kevin is now known as Guest56190 | 02:34 | |
*** anushkrishnamurt has quit IRC | 02:40 | |
*** muralia has joined #openstack-containers | 02:51 | |
*** muralia has quit IRC | 02:56 | |
*** jerrygb has joined #openstack-containers | 03:33 | |
*** jerrygb has quit IRC | 03:37 | |
*** Jeffrey4l has joined #openstack-containers | 03:44 | |
*** jwcroppe has quit IRC | 03:53 | |
*** sergmelikyan has joined #openstack-containers | 04:31 | |
*** sergmelikyan has quit IRC | 04:35 | |
*** jerrygb has joined #openstack-containers | 04:37 | |
*** muralia has joined #openstack-containers | 04:53 | |
*** hehx_ has quit IRC | 04:53 | |
*** muralia has quit IRC | 04:57 | |
openstackgerrit | Michael Liu proposed openstack/magnum: magnum support cni network driver https://review.openstack.org/391718 | 05:00 |
---|---|---|
*** tonanhngo has joined #openstack-containers | 05:04 | |
*** jerrygb has quit IRC | 05:04 | |
*** tonanhngo has quit IRC | 05:09 | |
*** chhavi has joined #openstack-containers | 05:25 | |
*** tonanhngo has joined #openstack-containers | 05:31 | |
*** Wenzhi has joined #openstack-containers | 05:38 | |
*** vimal has joined #openstack-containers | 05:43 | |
*** muralia has joined #openstack-containers | 05:54 | |
*** muralia has quit IRC | 05:59 | |
*** mjura has joined #openstack-containers | 06:26 | |
*** mjura_ has joined #openstack-containers | 06:26 | |
*** mjura has quit IRC | 06:26 | |
chhavi | I am trying to create a cluster using openstack-magnum, its failing for "os-collect-config:cfn unavailable" | 06:46 |
chhavi | checking further it looks cfn is default for os-collect-config, does that means the image should have the heat-cfn-tools | 06:47 |
*** muralia has joined #openstack-containers | 06:55 | |
*** belmoreira has joined #openstack-containers | 06:59 | |
openstackgerrit | Ryosuke Mizuno proposed openstack/magnum-ui: Add javascript tests for clusterCreateInfoController https://review.openstack.org/391727 | 06:59 |
chhavi | *heat-cfntools | 07:00 |
*** muralia has quit IRC | 07:00 | |
*** jerrygb has joined #openstack-containers | 07:02 | |
*** jerrygb has quit IRC | 07:08 | |
openstackgerrit | Michael Liu proposed openstack/magnum: magnum support cni network driver https://review.openstack.org/391718 | 07:12 |
*** Oku_OS-away is now known as Oku_OS | 07:12 | |
*** tonanhngo has quit IRC | 07:18 | |
*** yuanying has joined #openstack-containers | 07:27 | |
*** yuanying has quit IRC | 07:27 | |
*** chhavi has quit IRC | 07:33 | |
*** Jeffrey4l has quit IRC | 07:34 | |
*** tonanhngo has joined #openstack-containers | 07:41 | |
*** dimtruck is now known as zz_dimtruck | 07:43 | |
*** zz_dimtruck is now known as dimtruck | 07:43 | |
*** chhavi has joined #openstack-containers | 07:44 | |
*** tonanhngo has quit IRC | 07:46 | |
*** dimtruck is now known as zz_dimtruck | 07:53 | |
*** muralia has joined #openstack-containers | 07:56 | |
*** muralia has quit IRC | 08:00 | |
*** absubram has joined #openstack-containers | 08:02 | |
*** Oku_OS is now known as Oku_OS-away | 08:12 | |
*** Oku_OS-away is now known as Oku_OS | 08:13 | |
*** fragatina has joined #openstack-containers | 08:21 | |
*** Oku_OS is now known as Oku_OS-away | 08:21 | |
*** fragatina has quit IRC | 08:21 | |
*** fragatina has joined #openstack-containers | 08:22 | |
*** Wenzhi has quit IRC | 08:23 | |
*** fragatina has quit IRC | 08:36 | |
*** Oku_OS-away is now known as Oku_OS | 08:48 | |
*** kbyrne has quit IRC | 08:56 | |
*** muralia has joined #openstack-containers | 08:57 | |
*** muralia has quit IRC | 09:01 | |
*** chhavi has quit IRC | 09:29 | |
*** yuanying has joined #openstack-containers | 09:39 | |
*** tonanhngo has joined #openstack-containers | 09:42 | |
*** yuanying has quit IRC | 09:44 | |
*** tonanhngo has quit IRC | 09:47 | |
*** vimal has quit IRC | 09:47 | |
*** muralia has joined #openstack-containers | 09:57 | |
*** vimal has joined #openstack-containers | 09:57 | |
*** chhavi has joined #openstack-containers | 09:58 | |
*** muralia has quit IRC | 10:02 | |
strigazi | mjura_ jgrassler ping | 10:06 |
mjura_ | strigazi: I'm here | 10:07 |
mjura_ | good morning | 10:07 |
mjura_ | strigazi: how can I help you? | 10:07 |
strigazi | Good Morning, | 10:08 |
strigazi | Would suse be interested in providing a third party CI to run functional tests for the k8s_suse driver? | 10:09 |
strigazi | Do you know whom the magnum team should contact for this? | 10:09 |
strigazi | We plan to ask Canonical too, for ubuntu flavors | 10:10 |
strigazi | s/flavors/drivers | 10:10 |
mjura_ | strigazi: I think that we should be interested in | 10:12 |
mjura_ | I think we should escalate through few people | 10:13 |
*** apuimedo has joined #openstack-containers | 10:26 | |
openstackgerrit | Spyros Trigazis proposed openstack/magnum: Add user-domain in role creation https://review.openstack.org/391779 | 10:32 |
mjura_ | strigazi: ok, I've sent official email to our PM and CI guys | 10:40 |
mjura_ | strigazi: from my site this idea has +1 | 10:40 |
mjura_ | strigazi: but let's what will be discussion | 10:40 |
mjura_ | strigazi: I've added you to CC | 10:41 |
strigazi | mjura_, cool, Thanks! | 10:42 |
*** fragatina has joined #openstack-containers | 10:44 | |
*** jerrygb has joined #openstack-containers | 10:57 | |
*** anushkrishnamurt has joined #openstack-containers | 11:14 | |
*** chhavi has quit IRC | 11:20 | |
*** fragatina has quit IRC | 11:24 | |
*** chhavi has joined #openstack-containers | 11:34 | |
*** anushkrishnamurt has quit IRC | 11:36 | |
*** tonanhngo has joined #openstack-containers | 11:43 | |
*** tonanhngo has quit IRC | 11:48 | |
*** absubram has quit IRC | 11:53 | |
*** absubram_ has joined #openstack-containers | 11:53 | |
*** askb has quit IRC | 11:53 | |
*** muralia has joined #openstack-containers | 11:59 | |
*** jerrygb has quit IRC | 12:01 | |
*** muralia has quit IRC | 12:03 | |
*** tonanhngo has joined #openstack-containers | 12:17 | |
*** tonanhngo_ has joined #openstack-containers | 12:21 | |
*** tonanhngo has quit IRC | 12:22 | |
*** tonanhngo_ has quit IRC | 12:25 | |
*** fragatina has joined #openstack-containers | 12:29 | |
*** yuanying has joined #openstack-containers | 12:40 | |
*** Drago has joined #openstack-containers | 12:43 | |
*** yuanying has quit IRC | 12:45 | |
*** sergmelikyan has joined #openstack-containers | 12:46 | |
*** Drago has quit IRC | 12:47 | |
*** sergmelikyan has quit IRC | 12:50 | |
*** Drago has joined #openstack-containers | 12:52 | |
*** Drago has quit IRC | 12:52 | |
*** dave-mccowan has joined #openstack-containers | 12:53 | |
*** Drago has joined #openstack-containers | 12:55 | |
*** Drago1 has joined #openstack-containers | 12:57 | |
*** Drago has quit IRC | 13:00 | |
*** muralia has joined #openstack-containers | 13:00 | |
*** jerrygb has joined #openstack-containers | 13:03 | |
*** muralia has quit IRC | 13:04 | |
*** jerrygb has quit IRC | 13:07 | |
*** tonanhngo has joined #openstack-containers | 13:08 | |
*** jerrygb has joined #openstack-containers | 13:11 | |
*** jerrygb has quit IRC | 13:11 | |
*** Drago1 has quit IRC | 13:11 | |
*** Drago has joined #openstack-containers | 13:12 | |
*** tonanhngo has quit IRC | 13:12 | |
*** Drago has quit IRC | 13:14 | |
*** absubram_ has quit IRC | 13:14 | |
*** jmckind has joined #openstack-containers | 13:19 | |
*** jwcroppe has joined #openstack-containers | 13:32 | |
openstackgerrit | Spyros Trigazis proposed openstack/magnum: WIP: Make cinder volume optional https://review.openstack.org/391830 | 13:35 |
*** Drago has joined #openstack-containers | 13:37 | |
*** Drago1 has joined #openstack-containers | 13:41 | |
*** Drago has quit IRC | 13:42 | |
*** Drago1 has quit IRC | 13:45 | |
*** sergmelikyan has joined #openstack-containers | 13:56 | |
*** absubram has joined #openstack-containers | 13:56 | |
*** Drago1 has joined #openstack-containers | 14:00 | |
*** muralia has joined #openstack-containers | 14:00 | |
*** muralia has quit IRC | 14:05 | |
*** jerrygb has joined #openstack-containers | 14:08 | |
*** zz_dimtruck is now known as dimtruck | 14:09 | |
*** sergmelikyan has quit IRC | 14:13 | |
*** sergmelikyan has joined #openstack-containers | 14:13 | |
*** chris_hultin|AWA is now known as chris_hultin | 14:14 | |
*** mtanino has joined #openstack-containers | 14:15 | |
*** sergmelikyan has quit IRC | 14:19 | |
*** briancurtin has quit IRC | 14:19 | |
*** muralia has joined #openstack-containers | 14:20 | |
*** mtanino has quit IRC | 14:20 | |
*** briancurtin has joined #openstack-containers | 14:22 | |
*** clsacramento has joined #openstack-containers | 14:25 | |
*** sergmelikyan has joined #openstack-containers | 14:30 | |
*** JoseMello has joined #openstack-containers | 14:33 | |
*** randallburt has joined #openstack-containers | 14:34 | |
*** sergmelikyan has quit IRC | 14:35 | |
*** jmckind_ has joined #openstack-containers | 14:36 | |
*** jmckind has quit IRC | 14:38 | |
*** randallburt has quit IRC | 14:38 | |
*** randallburt has joined #openstack-containers | 14:41 | |
*** fragatina has quit IRC | 14:42 | |
*** mjura_ has quit IRC | 14:53 | |
*** chhavi has quit IRC | 14:54 | |
*** vimal has quit IRC | 14:55 | |
*** syed_ has joined #openstack-containers | 14:58 | |
*** Drago1 has quit IRC | 15:02 | |
*** Drago has joined #openstack-containers | 15:03 | |
*** jerrygb has quit IRC | 15:07 | |
*** Drago has quit IRC | 15:08 | |
*** jerrygb has joined #openstack-containers | 15:08 | |
*** tonanhngo has joined #openstack-containers | 15:08 | |
*** Drago has joined #openstack-containers | 15:12 | |
*** tonanhngo has quit IRC | 15:13 | |
*** tonanhngo has joined #openstack-containers | 15:19 | |
*** chhavi has joined #openstack-containers | 15:19 | |
*** mtanino has joined #openstack-containers | 15:22 | |
*** tonanhngo has quit IRC | 15:23 | |
*** Jeffrey4l has joined #openstack-containers | 15:38 | |
*** Jeffrey4l has quit IRC | 15:39 | |
*** sergmelikyan has joined #openstack-containers | 15:40 | |
*** Jeffrey4l has joined #openstack-containers | 15:40 | |
*** Jeffrey4l has quit IRC | 15:41 | |
*** apuimedo has quit IRC | 15:42 | |
*** Jeffrey4l has joined #openstack-containers | 15:42 | |
*** tonanhngo has joined #openstack-containers | 15:49 | |
*** tonanhngo has quit IRC | 15:53 | |
Drago | rochaporto: ping | 15:57 |
*** srwilkers has joined #openstack-containers | 15:57 | |
*** swatson_ has joined #openstack-containers | 15:57 | |
Drago | strigazi: ping | 15:59 |
*** chetna has joined #openstack-containers | 16:05 | |
*** absubram has quit IRC | 16:12 | |
*** AlexeyAbashkin has quit IRC | 16:13 | |
*** AlexeyAbashkin has joined #openstack-containers | 16:15 | |
*** jmckind_ has quit IRC | 16:17 | |
*** anush has joined #openstack-containers | 16:19 | |
*** chetna has quit IRC | 16:19 | |
*** chetna has joined #openstack-containers | 16:20 | |
strigazi | hey Drago | 16:20 |
Drago | Hey strigazi, about the NodeGroup discussion at the summit | 16:20 |
Drago | strigazi: Do you remember why rochaporto said it wasn't acceptable to not have a reference to the CT in the Cluster? | 16:21 |
Drago | strigazi: For some reason, simply comparing all of the important attributes in a Cluster to a supported ClusterTemplate wasn't sufficient to know if a Cluster was "supported" | 16:22 |
strigazi | Drago, I think it was for the upgrades, to have the increasing counter | 16:22 |
Drago | jvgrant ^ | 16:23 |
jvgrant | strigazi, counter of current version? not sure i understand | 16:24 |
strigazi | be back in 10' | 16:25 |
Drago | jvgrant: The idea is that as the CT gets updated, it can have an incrementing version along with it | 16:25 |
*** jmckind has joined #openstack-containers | 16:25 | |
*** dimtruck is now known as zz_dimtruck | 16:25 | |
*** zz_dimtruck is now known as dimtruck | 16:26 | |
jvgrant | ok, so it is about wanting to maintain versioning information on cluster creation | 16:30 |
Drago | jvgrant: I'm not sure anymore. I feel like I'm speculating at this point. The upgrade part didn't come in until near the end of the conversation we had about it | 16:31 |
jvgrant | Drago: ok, probably need to have some further discussion on that use case then. | 16:32 |
*** rafael__ has joined #openstack-containers | 16:34 | |
*** kbyrne has joined #openstack-containers | 16:41 | |
*** absubram has joined #openstack-containers | 16:41 | |
Drago | strigazi: You there? | 16:46 |
*** tonanhngo has joined #openstack-containers | 16:50 | |
*** david-lyle has joined #openstack-containers | 16:51 | |
*** Jeffrey4l has quit IRC | 16:52 | |
*** tonanhngo has quit IRC | 16:54 | |
*** tonanhngo has joined #openstack-containers | 16:58 | |
*** belmoreira has quit IRC | 16:58 | |
*** tonanhngo has quit IRC | 17:02 | |
*** srwilkers has quit IRC | 17:09 | |
Drago | jvgrant: Thinking more about it, I think the upgrade story wasn't any more complicated than that. When you update the CT, it increments its version, so you can then query Clusters created from foo_ct v1, v2, etc | 17:11 |
jvgrant | Drago: ok so there would be a new version field in the cluster for CT version? or are we going back to separating the objects into cluster and clustertemplate again? | 17:13 |
jvgrant | i guess if we just add it to the cluster object then the version is set when the cluster create happens. | 17:15 |
jvgrant | i think that works fine if we just add a CT_name attribute and CT_version attribute. when CTs are created/updated they are changed and in the case of a cluster create they are filled in from the CT(if present) or left unset to indicate it was not created from a CT | 17:17 |
Drago | jvgrant: Since we're using a relational DB, I'd think we would add the immutable CT's FK to the Cluster and the the immutable CT would have the version in it | 17:17 |
jvgrant | Drago: that works as well, but would end up with a lot more db entries | 17:17 |
*** tonanhngo has joined #openstack-containers | 17:18 | |
Drago | Workflow: Create a CT (version is 1). Update it (version bumps to 2). Create a Cluster - immutable CT gets created that, naturally, has v2 and set the FK on the Cluster to it. Create another cluster - uses previously-created immutable CT | 17:19 |
Drago | jvgrant: CTs should still be prototypes and updatable | 17:19 |
Drago | jvgrant: And Clusters should still have all attributes, since who knows which ones will be updatable. I think the immutable CT should be there just to be reference-able. | 17:20 |
jvgrant | Drago: ok, makes sense | 17:21 |
jvgrant | Drago: can you delete CTs then? | 17:21 |
*** tonanhngo has quit IRC | 17:22 | |
Drago | jvgrant: Good question. Should be able to, not sure how it'd work with what we're talking about. | 17:22 |
jvgrant | Drago: yeah, seems a little complicated with now versioned CTs | 17:23 |
Drago | jvgrant: Might want to do something like have the client prompt to make sure the user really wants to delete the CT and then delete all versioning information | 17:24 |
jvgrant | Drago: what if it is referenced in a cluster? | 17:24 |
*** srwilkers has joined #openstack-containers | 17:25 | |
jvgrant | Drago: maybe the reference in the cluster should be more informational. Just the name and version. They can look it up but it might not be there | 17:25 |
Drago | jvgrant: It wouldn't be an issue if the relation between immutable CTs and Clusters was many-to-many. Reason for that is you can create a Cluster from multiple CTs right now, and I like that feature | 17:25 |
Drago | jvgrant: Not sure I'm a fan of that | 17:26 |
jvgrant | Drago: it is up to the operator to maintain the versions if they wish. A updated CT version is really just a copy of the original with a new version number and the changes | 17:27 |
*** jwcroppe has quit IRC | 17:27 | |
jvgrant | Drago: i'm feeling like we are going back to required CTs that are separate objects really | 17:27 |
*** jwcroppe has joined #openstack-containers | 17:28 | |
strigazi | Drago, sorry there was an issue with upgrade we did | 17:29 |
Drago | jvgrant: I don't see how. What we're talking about is the same as we have in the spec except for the addition of a reference to an immutable, informational copy of the CT used | 17:29 |
Drago | jvgrant: Once the cluster is created, the CT still has nothing to do with it, functionally | 17:30 |
Drago | strigazi: No problem | 17:30 |
jvgrant | Drago: i'm ok with that if that is all it is. But if it is also making it so we now require the original CT to not be deleted, etc.. then i think that is more complicated | 17:31 |
*** jwcroppe_ has joined #openstack-containers | 17:31 | |
jvgrant | Drago: if we are just storing a snapshot then i don't see a problem | 17:31 |
Drago | jvgrant: I think we need to figure out what we want to do about the version information upon deletion, but all the other operations are the same (besides incrementing the CT version) | 17:32 |
*** jwcroppe has quit IRC | 17:32 | |
strigazi | what about the NGTs? | 17:33 |
strigazi | Drago, jvgrant ^^ | 17:33 |
jvgrant | they would follow the same rules as the CTs | 17:33 |
Drago | strigazi: I don't know what the edge cases are for NGTs, but yes, the same rules should apply | 17:33 |
*** yuanying has joined #openstack-containers | 17:34 | |
jvgrant | each cluster would have references to any templates it used to create itself and the contained nodegroups as well | 17:34 |
strigazi | Drago, jvgrant, the reasoning behind version is that we want to track where a cluster came from | 17:34 |
Drago | strigazi: So a query like "give me clusters from my_ct v1" is what you desire? | 17:35 |
jvgrant | strigazi: all the info of it or just something that could be used to track it? Meaning a snapshot of the CT used or just a name/version info that could be used to look it up? | 17:36 |
strigazi | If we don't allow deletions name/version should be enough | 17:37 |
*** yuanying has quit IRC | 17:38 | |
strigazi | To manage clusters in bulk you need to have them grouped | 17:38 |
strigazi | These clusters come from CT_x, do action action_A on them | 17:39 |
jvgrant | i'm assuming you also want to be able to block any overrides during cluster creation then from the CT? | 17:39 |
strigazi | jvgrant, Drago, make sense? | 17:39 |
Drago | strigazi: So the user could create a CT of the same name and it would come up in your query? | 17:40 |
strigazi | We talked about that too, locking should be applied when needed | 17:40 |
strigazi | Drago, even if it comes up we shouldn't care, not sure how this can be handled | 17:42 |
*** jwcroppe_ has quit IRC | 17:42 | |
jvgrant | this seems like it could be handled through admin settings limiting the creation/deletion of CTs. then the references with name/version would be enough for the grouping | 17:43 |
chhavi | jvgrant, Drago: I am trying to create a cluster, its failing in os-collect-config for cfn ssl verification failure. | 17:43 |
strigazi | Drago jvgrant, Do you still have concerns why we want to keep track of CTs? | 17:44 |
chhavi | its hitting this bug: https://bugs.launchpad.net/heat/+bug/1482510 | 17:45 |
openstack | Launchpad bug 1482510 in heat "OS::Heat::SoftwareDeployment failed due SSL certificate verification error" [Medium,In progress] - Assigned to Steve Baker (steve-stevebaker) | 17:45 |
Drago | strigazi, jvgrant: Seems janky to have a feature like that just to not have references to things in the data model | 17:45 |
jvgrant | strigazi: i think i understand the use case now a little better. Just want to make sure we don't over engineer a solution that some small changes on top could fix | 17:45 |
strigazi | "some small changes on top could fix" what do you mean? | 17:45 |
Drago | chhavi: Using the Mesos driver? | 17:46 |
Drago | I think that's the only one with os-collect-config | 17:47 |
chhavi | yes i am using ubuntu mesos image | 17:47 |
chhavi | on ppc64le | 17:47 |
jvgrant | strigazi: meaning instead of implementing a full versioning functionality for CTs and storing an additional object for each template on creation, could we get the same use case by simple references and other admin options like limiting the creation/deletion when desired | 17:47 |
Drago | chhavi: If you want, you can configure Heat to use a different software config transport by default. It uses the CFN API by default, and that's probably not what you really want. Change it to use Heat, or swift or zaqar if you want | 17:48 |
jvgrant | strigazi, Drago: we might still need to store the additional info, i'm just playing devil's advocate here to make sure we aren't over complicating this | 17:49 |
Drago | jvgrant: From what I understand, strigazi doesn't want to stop users from creating their own CTs. | 17:49 |
*** dave-mccowan has quit IRC | 17:49 | |
strigazi | Drago, ^^ yes | 17:49 |
strigazi | but to manage cluster, specifically for upgrade, we don't want to allow unilimited combinations of attributes as in the proposed model | 17:50 |
strigazi | but to manage clusters, specifically for upgrades, we don't want to allow unilimited combinations of attributes as in the proposed model | 17:50 |
Drago | strigazi: Right, locked attributes covers that one though | 17:51 |
Drago | Just not the tracking | 17:51 |
jvgrant | strigazi: so you want to know which template(s) were used to create any CT | 17:51 |
strigazi | Drago, you need tracking | 17:51 |
Drago | chhavi: https://github.com/openstack/heat/blob/master/heat/common/config.py#L195-L214 | 17:51 |
chhavi | Drago: is the cfn-api not functional for mesos | 17:51 |
jvgrant | strigazi: any cluster not CT | 17:51 |
strigazi | Drago, we won't always upgrade all clusters in one go, some will upgrade now some later | 17:52 |
strigazi | jvgrant, yes? | 17:52 |
strigazi | jvgrant, yes | 17:52 |
Drago | chhavi: Mesos has nothing to do with it. It's probably just a bug with softwareconfig and the CFN API. I would just switch away from the CFN API anyway, which would also solve your problem | 17:53 |
*** tushark has joined #openstack-containers | 17:53 | |
chhavi | Drago: ok i will give a try to use heat, thanks | 17:53 |
jvgrant | strigazi: what if you deleted the CT? | 17:53 |
Drago | strigazi: And how will you track the ones you already did vs later? | 17:54 |
strigazi | You don't if it's referenced | 17:54 |
strigazi | upgraded clusters will have a bumbed version. | 17:55 |
Drago | strigazi: In the model we're talking about, all CTs that had a Cluster created from them would be "referenced" so I think that's a tough policy to sell | 17:55 |
strigazi | Drago, that can be true only in some cases I guess | 17:56 |
strigazi | I understand the desire to create clusters at will but we must also track them | 17:56 |
Drago | strigazi: The main reason is that you can update the CT independently. It makes sense to keep the *versioned* (immutable) CT around, but not the one you can create Clusters from | 17:56 |
jvgrant | strigazi, Drago: what if we don't allow CT updates? You have to create a new one. It can be based on the old, but has a new name | 17:57 |
Drago | jvgrant: short answer: No | 17:57 |
jvgrant | strigazi,Drago: and then we give an option to block CT deletes | 17:58 |
Drago | jvgrant: long answer: Noooooo | 17:58 |
strigazi | Drago, what do you propose then? | 17:58 |
strigazi | Solve the following problem | 17:59 |
Drago | jvgrant: Why a new name? | 17:59 |
jvgrant | strigazi, Drago: it seems like the main use case we are talking about is tracking and labeling. Know what template a cluster/nodegroup came from. | 17:59 |
Drago | jvgrant: I guess honestly, the system we're imagining so far is kind of like what you said, except the immutable CT is created when you create the cluster, while the "normal" CT is mutable | 18:00 |
jvgrant | if i have MesosCT_v1 and then i want an update cluster i just create it referencing MesosCT_v1 and name it MesosCT_v2 | 18:01 |
jvgrant | since as an admin operator i blocked CT deletion of "public" CTs | 18:01 |
Drago | jvgrant: I think versioning would be nicer than name munging | 18:01 |
Drago | For the same effect | 18:01 |
jvgrant | all i need in the cluster is a reference to the name | 18:01 |
Drago | jvgrant: And with names you have the danger of collisioins | 18:02 |
Drago | *collisions | 18:02 |
jvgrant | the problem is the reference is just a snap shot and if the actually CT was deleted or changed it doesn't really tell the operator anything accept what was used | 18:02 |
jvgrant | Drago: yeah, it could be a version field instead of name | 18:03 |
strigazi | If you snapshot inside the cluster object you can't query I guess | 18:03 |
Drago | jvgrant: I kind of see your point with making CTs immutable, but I wouldn't want a ton of CTs showing up in the UI | 18:04 |
strigazi | Think of the two types of clusters | 18:04 |
strigazi | supported by the ops team | 18:04 |
strigazi | and the ones created by users | 18:04 |
strigazi | we want to manage the supported clusters | 18:04 |
strigazi | Total freedom on how to create clusters and be able to sell it | 18:05 |
strigazi | applies to the second type | 18:05 |
strigazi | We also want to sell | 18:05 |
Drago | jvgrant: What if we had the CTs all reference a "parent" thing that would help track the different CT versions, since they'd all be linked to it. In the UI we can expose only the most recent CT version found | 18:05 |
strigazi | run your prodution service here and you'll get the new version without downtime | 18:05 |
jvgrant | do controls over "public"(for ops team) and "private" CT help that? Limit what can be done to "public" CTs | 18:06 |
strigazi | I think it helps | 18:07 |
Drago | jvgrant: To be more clear, when create a new CT, the CT parent (for lack of a better name) would be created as well. When you modify a CT, it creates a new CT of the next version that also references this parent | 18:07 |
strigazi | this the only way to do versions | 18:07 |
strigazi | this is the only way to do versions | 18:07 |
Drago | If you wanted to delete the CT, you'd delete that new one and the parent would stick around | 18:08 |
*** achanda has joined #openstack-containers | 18:08 | |
Drago | jvgrant, strigazi: Up until now we've been talking about the versions of a CT referencing the original CT. I'm saying switch that to something that only has the purpose of "gathering" the CT versions together | 18:09 |
jvgrant | Drago: we are now talking about a new versioning feature on top of templates | 18:09 |
achanda | eliqiao: is this feature complete now? https://blueprints.launchpad.net/magnum/+spec/support-insecure-registry | 18:10 |
Drago | jvgrant: Well, yeah, the conversation has been moving towards that | 18:10 |
jvgrant | Drago: i like that better than trying to imbed too much into the templates | 18:10 |
jvgrant | and clusters | 18:10 |
*** tonanhngo has joined #openstack-containers | 18:12 | |
strigazi | I'm leaving guys, it's late here. I'll do my best to describe the upgrades in detail | 18:12 |
Drago | strigazi: See ya | 18:12 |
jvgrant | strigazi: thanks for your help | 18:12 |
Drago | jvgrant: Creating a new CT entry for every modification feels like overkill. Unless there are good reasons for tracking every modification, I'd rather track the modifications that actually make it into a cluster | 18:13 |
Drago | jvgrant: So, have a "real" CT that gets modified, and create a *single* immutable CT for clusters that use a given version | 18:14 |
jvgrant | Drago: so the immutable CT is create on CT modification, not on cluster create | 18:15 |
*** chhavi has quit IRC | 18:15 | |
Drago | jvgrant: I'm saying I'd rather have it on cluster create | 18:15 |
*** tonanhngo has quit IRC | 18:17 | |
jvgrant | Drago: that doesn't really help with the tracking issue does it? I guess i don't understand the need for knowing all the details of what was used. This seems more like labeling the cluster so it can be tracked vs needing to know all the values taht were used | 18:17 |
jvgrant | Drago: in the use cases mentioned it only matters if there is a valid public CT that it can match to | 18:18 |
jvgrant | Drago: if that CT is gone, then the info stored doesnt' really mean much | 18:18 |
Drago | jvgrant: I'm not sure what you're responding to, because that seems orthogonal to what I was saying | 18:19 |
Drago | jvgrant: I think we should vidyo | 18:19 |
jvgrant | Drago: sure, send me a link. we might be going different directions on this | 18:19 |
*** dave-mccowan has joined #openstack-containers | 18:19 | |
*** jmckind has quit IRC | 18:21 | |
*** tonanhngo has joined #openstack-containers | 18:22 | |
*** jmckind has joined #openstack-containers | 18:22 | |
*** jmckind_ has joined #openstack-containers | 18:25 | |
*** tonanhngo has quit IRC | 18:26 | |
*** jmckind has quit IRC | 18:29 | |
*** EricGonczer_ has joined #openstack-containers | 18:36 | |
*** EricGonczer_ has quit IRC | 18:37 | |
tushark | Hi, I have kubernetes cluster with TLS enabled and setup with SkyDns, but the kubedns fails to talk to the api-server. | 18:40 |
tushark | E1031 17:08:18.998416 1 reflector.go:214] pkg/dns/dns.go:155: Failed to list *api.Endpoints: Get https://10.240.126.7:6443/api/v1/endpoints?resourceVersion=0: x509: failed to load system roots and no roots provided | 18:41 |
tushark | E1031 17:08:18.998703 1 reflector.go:214] pkg/dns/dns.go:156: Failed to list *api.Service: Get https://10.240.126.7:6443/api/v1/services?resourceVersion=0: x509: failed to load system roots and no roots provided | 18:41 |
tushark | Although I'm not configuring secrets for the service account, is this causing the above failure? | 18:41 |
*** chetna has quit IRC | 18:45 | |
*** chetna has joined #openstack-containers | 18:45 | |
*** fragatina has joined #openstack-containers | 18:53 | |
*** fragatina has quit IRC | 19:13 | |
*** tonanhngo has joined #openstack-containers | 19:14 | |
*** tonanhngo has quit IRC | 19:18 | |
*** randallburt has quit IRC | 19:23 | |
*** tonanhngo has joined #openstack-containers | 19:30 | |
*** apuimedo has joined #openstack-containers | 19:30 | |
*** tonanhngo has quit IRC | 19:34 | |
*** apuimedo has quit IRC | 19:37 | |
*** tonanhngo has joined #openstack-containers | 19:40 | |
openstackgerrit | Vijendar Komalla proposed openstack/magnum: [WIP]Admin api to provide cluster stats https://review.openstack.org/391301 | 19:43 |
*** tonanhngo_ has joined #openstack-containers | 19:44 | |
*** tonanhngo has quit IRC | 19:44 | |
*** tonanhngo_ has quit IRC | 19:48 | |
*** tonanhngo has joined #openstack-containers | 19:49 | |
*** anush has quit IRC | 19:52 | |
*** tonanhngo has quit IRC | 19:53 | |
*** mikelekkas has joined #openstack-containers | 19:58 | |
*** mikelekkas has quit IRC | 19:59 | |
*** Drago has quit IRC | 20:03 | |
*** muralia has quit IRC | 20:09 | |
*** muralia has joined #openstack-containers | 20:09 | |
*** muralia has quit IRC | 20:14 | |
*** dave-mccowan has quit IRC | 20:15 | |
*** achanda has quit IRC | 20:16 | |
*** DanyC has joined #openstack-containers | 20:19 | |
*** drago has joined #openstack-containers | 20:20 | |
*** JoseMello has quit IRC | 20:22 | |
*** drago has quit IRC | 20:26 | |
*** DanyC has left #openstack-containers | 20:27 | |
*** drago has joined #openstack-containers | 20:28 | |
*** Drago5 has joined #openstack-containers | 20:33 | |
*** drago has quit IRC | 20:38 | |
*** Drago5 is now known as Drago | 20:40 | |
*** jwcroppe has joined #openstack-containers | 20:44 | |
*** dave-mccowan has joined #openstack-containers | 20:45 | |
*** jerrygb__ has joined #openstack-containers | 20:48 | |
*** jerrygb has quit IRC | 20:49 | |
*** srwilkers has quit IRC | 20:51 | |
*** jerrygb__ has quit IRC | 20:52 | |
*** jwcroppe has quit IRC | 20:55 | |
*** chris_hultin is now known as chris_hultin|AWA | 21:01 | |
*** DanyC has joined #openstack-containers | 21:06 | |
*** DanyC has quit IRC | 21:11 | |
*** apuimedo has joined #openstack-containers | 21:14 | |
*** openstackgerrit has quit IRC | 21:18 | |
*** openstackgerrit has joined #openstack-containers | 21:19 | |
*** vijendar has joined #openstack-containers | 21:28 | |
*** askb has joined #openstack-containers | 21:29 | |
*** dave-mccowan has quit IRC | 21:38 | |
*** tonanhngo has joined #openstack-containers | 21:49 | |
*** absubram has quit IRC | 21:50 | |
*** vijendar has quit IRC | 21:51 | |
*** vijendar has joined #openstack-containers | 21:54 | |
*** tonanhngo has quit IRC | 21:54 | |
*** jerrygb has joined #openstack-containers | 21:59 | |
*** jerrygb has quit IRC | 22:04 | |
*** sergmelikyan has quit IRC | 22:19 | |
*** vije5904 has joined #openstack-containers | 22:23 | |
*** vije5904 has left #openstack-containers | 22:23 | |
*** vijendar has quit IRC | 22:24 | |
*** vijendar has joined #openstack-containers | 22:25 | |
*** tonanhngo has joined #openstack-containers | 22:28 | |
*** sergmelikyan has joined #openstack-containers | 22:29 | |
*** tonanhngo has quit IRC | 22:33 | |
*** vijendar has quit IRC | 22:37 | |
*** sergmelikyan has quit IRC | 22:39 | |
*** jerrygb has joined #openstack-containers | 22:42 | |
*** jerrygb_ has joined #openstack-containers | 22:45 | |
*** chetna has quit IRC | 22:48 | |
*** jerrygb has quit IRC | 22:48 | |
*** tonanhngo has joined #openstack-containers | 22:53 | |
*** vijendar has joined #openstack-containers | 22:55 | |
*** apuimedo has quit IRC | 22:56 | |
*** muralia has joined #openstack-containers | 22:57 | |
*** tonanhngo has quit IRC | 22:57 | |
*** vijendar1 has joined #openstack-containers | 22:58 | |
*** srwilkers has joined #openstack-containers | 22:59 | |
openstackgerrit | Vijendar Komalla proposed openstack/magnum: Admin api to provide cluster stats https://review.openstack.org/391301 | 22:59 |
*** muralia has quit IRC | 23:00 | |
*** muralia has joined #openstack-containers | 23:00 | |
*** tonanhngo has joined #openstack-containers | 23:01 | |
*** rafael__ has quit IRC | 23:05 | |
*** vijendar has quit IRC | 23:05 | |
*** vijendar1 has quit IRC | 23:05 | |
*** vijendar has joined #openstack-containers | 23:06 | |
*** tonanhngo has quit IRC | 23:06 | |
*** vijendar has quit IRC | 23:07 | |
*** apuimedo has joined #openstack-containers | 23:07 | |
*** vijendar has joined #openstack-containers | 23:07 | |
*** chetna has joined #openstack-containers | 23:07 | |
*** vijendar_ has joined #openstack-containers | 23:08 | |
*** clsacramento has quit IRC | 23:25 | |
*** vijenda__ has joined #openstack-containers | 23:26 | |
*** clsacramento has joined #openstack-containers | 23:26 | |
openstackgerrit | Stephen Watson proposed openstack/magnum: [WIP] Adds support for non-ID suffix https://review.openstack.org/389811 | 23:27 |
*** vijendar_ has quit IRC | 23:28 | |
*** vijendar1 has joined #openstack-containers | 23:28 | |
*** vijendar has quit IRC | 23:28 | |
*** apuimedo has quit IRC | 23:32 | |
*** Drago has quit IRC | 23:32 | |
*** yuanying has joined #openstack-containers | 23:38 | |
*** tonanhngo has joined #openstack-containers | 23:42 | |
*** apuimedo has joined #openstack-containers | 23:43 | |
*** vijendar has joined #openstack-containers | 23:44 | |
*** vijendar1 has quit IRC | 23:45 | |
*** vijenda__ has quit IRC | 23:45 | |
*** tonanhngo has quit IRC | 23:48 | |
*** vijendar1 has joined #openstack-containers | 23:49 | |
*** chetna has quit IRC | 23:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!