| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-06-22 | |||
| 16:40:59 | gibi | if we never mix, then I'm fine with the different RC | |
| 16:41:00 | bauzas | gibi: because inventories would be managed by different services | |
| 16:41:03 | sean-k-mooney | elodilles: cool just ping us and we can confirm if tis the same isssue or not | |
| 16:41:08 | sean-k-mooney | elodilles: hopefully at least | |
| 16:41:15 | bauzas | the consumption would be identical tho | |
| 16:41:20 | elodilles | sean-k-mooney: sure, thanks! :) | |
| 16:41:37 | sean-k-mooney | bauzas: we dont need different RC classes to mix | |
| 16:41:48 | bauzas | sean-k-mooney: yup, the other way | |
| 16:42:02 | sean-k-mooney | we dont need them to isolate ither | |
| 16:42:40 | gibi | sean-k-mooney: if you even want to support give-me-a-vgpu-i-dont-care-if-nova-or-cyborg-managed then we need a common pool of resource and therefore a common RC | |
| 16:42:42 | bauzas | sean-k-mooneybut I'm not sure marking an inventory by a trait for knowing about the owner is the best | |
| 16:42:43 | sean-k-mooney | a trait + RC can do eveything 2 RC classes can do since we do not allow two services to create inventories on teh same RP | |
| 16:43:19 | sean-k-mooney | gibi: right if we want to supprot that your are corect a common resouce class woudl be required | |
| 16:43:34 | gibi | if we don't what that then no need for a common RC | |
| 16:43:37 | bauzas | but I think we're missing some logic | |
| 16:44:05 | sean-k-mooney | bauzas: its marking the RP not the inventory | |
| 16:44:11 | bauzas | either we wanna mix, and then we don't need to know which inventory was created by who | |
| 16:44:32 | sean-k-mooney | bauzas: we do | |
| 16:44:33 | bauzas | or, we just make resources alongside | |
| 16:44:48 | sean-k-mooney | we do not allow 2 services to modify the same resouce provider | |
| 16:44:48 | bauzas | and then we don't need the RCs to be the same | |
| 16:45:00 | bauzas | really ? | |
| 16:45:07 | bauzas | I don't think we want to avoid this | |
| 16:45:09 | sean-k-mooney | so fi we have ownert traits or resocue classes we are fine | |
| 16:45:26 | bauzas | the generation bit is even for distributed servcies | |
| 16:45:36 | gmann | sean-k-mooney: stephenfin gibi can you point me to bug where server group validation did not work. I can see test validating the scema also - https://github.com/openstack/nova/blob/master/nova/tests/unit/api/openstack/compute/test_server_groups.py#L459 | |
| 16:45:50 | sean-k-mooney | bauzas: we dont allow it becaue the virt driver will replace the traits on the RP | |
| 16:46:22 | stephenfin | gmann: not the exact bug but https://storyboard.openstack.org/#!/story/2008975 and https://github.com/openstack/python-openstackclient/commit/ab0b1fe885ee0a210a58008b631521025be7f3eb | |
| 16:46:27 | stephenfin | I just posted to openstack-discuss about this | |
| 16:46:36 | sean-k-mooney | bauzas: and it will update the inventoris based on its local view | |
| 16:46:42 | gibi | I have to drop of now. I will read back tomorrow | |
| 16:46:46 | gmann | stephenfin: ok, thanks checking | |
| 16:47:06 | bauzas | sean-k-mooney: how placement know which service owns which RP ? | |
| 16:47:18 | sean-k-mooney | bauzas: it does not and it does not enforce it | |
| 16:47:32 | bauzas | sean-k-mooney: that's my point | |
| 16:47:34 | sean-k-mooney | we enforce this today by tellint neutorn and cybrog they may not modify our RPs | |
| 16:47:49 | bauzas | we're blindly recreating a RP | |
| 16:47:51 | gmann | stephenfin: so this is about removing the validation not that current validation did not work right? | |
| 16:47:54 | sean-k-mooney | bauzas: yep | |
| 16:48:22 | stephenfin | gmann: the bug is about removing the API validation from the client. sean-k-mooney is saying that the server validation isn't working though | |
| 16:48:36 | bauzas | sean-k-mooney: that's my point, some service owns some RP | |
| 16:48:36 | sean-k-mooney | well not quite | |
| 16:48:46 | bauzas | it's just nova/placement which doesn't support it | |
| 16:49:01 | sean-k-mooney | stephenfin: gmann im saying that the customer reported it worked and i cant see how the server validation would not block this | |
| 16:49:02 | bauzas | so | |
| 16:49:12 | bauzas | say cyborg creates some RPs | |
| 16:49:17 | bauzas | and nova too | |
| 16:49:26 | bauzas | both are different but have the same RCs | |
| 16:49:36 | spatel | sean-k-mooney hey! had quick question related Cellv2 design, does neutron can be scale using cellv2 or just nova? | |
| 16:49:42 | gmann | sean-k-mooney: stephenfin yeah it is rejected from server side in v2.1 at least (v2 old code I am not sure) | |
| 16:50:04 | bauzas | sean-k-mooney: then why should we have problems if a flavor is asking for some vGPUs ? | |
| 16:50:08 | sean-k-mooney | gmann: it shoudl be rejected in 2.0 based on the validation | |
| 16:50:23 | bauzas | sean-k-mooney: this would eventually go to the libvirt driver | |
| 16:50:41 | sean-k-mooney | bauzas: right now cybrog neutron and other prject cannot use any standard RC used by nova | |
| 16:50:41 | bauzas | which would know whether the allocated RP is owned by it or cyboth | |
| 16:50:44 | gmann | sean-k-mooney: v2 code path was different until we merged it. I am not sure if user using that much old nova | |
| 16:51:23 | sean-k-mooney | gmann: ah good point it was working for them on osp 13 which is queens and not workin on osp 16 whic is train | |
| 16:51:35 | gmann | sean-k-mooney: and in old v2 code path I think we did not had schema validation. its v2.1 introducing the JSON schema validation in more strict way | |
| 16:52:07 | sean-k-mooney | bauzas: perhaps we could just have a google meet call | |
| 16:52:31 | bauzas | sean-k-mooney: unfortunately, I need to stop in 5 mins | |
| 16:52:43 | bauzas | gosh I miss PTGs | |
| 16:52:49 | bauzas | and f2f meetings | |
| 16:52:55 | bauzas | and whiteboards | |
| 16:53:13 | sean-k-mooney | bauzas: we can talk tabout it tomorrow then but tl;dr; today no too openstack service can use the same RC class and no two openstack service can share the same RP | |
| 16:53:14 | bauzas | I could even show source code | |
| 16:53:35 | sean-k-mooney | bauzas: the ownwer triat was a way to allow reuse of stanard RCs across services | |
| 16:53:43 | bauzas | sean-k-mooney: based on your first assumption, that means cyborg can't provide VGPU resources ? | |
| 16:53:59 | bauzas | sean-k-mooney: gosh, this sounds hacky | |
| 16:54:05 | sean-k-mooney | correct which is why the ownwer trait was intoduced in the spec | |
| 16:54:21 | gmann | sean-k-mooney: Mitaka was last with legacy v2 code and since newton we started single code path for v2 and v2.1 which has these schema validation | |
| 16:54:36 | sean-k-mooney | without an owner of the resouce provider in some form resouce classes cant be shared | |
| 16:54:51 | bauzas | sean-k-mooney: honestly, I'd prefer us to have a way to know about which service owns a RP with by default be 'nova' | |
| 16:55:06 | bauzas | meaning another attribute of a RP | |
| 16:55:10 | sean-k-mooney | bauzas: well that is exactly what the ownwer traits were ment to be | |
| 16:55:26 | sean-k-mooney | bauzas: since peole did not want a dedicated value on the RP | |
| 16:55:29 | bauzas | the owner trait is just a hack because we wanna make it round | |
| 16:55:49 | sean-k-mooney | bauzas: we proposed added ownwer before and peopel objected | |
| 16:55:58 | bauzas | sean-k-mooney: then I don't guess why we need to explicitely mark a nova trait | |
| 16:56:05 | sean-k-mooney | we can certenly do that again but the default shoudl nto be nova | |
| 16:56:15 | gmann | sean-k-mooney: stephenfin ohk its modified code he mentioned in github discussion so upstream code is all good | |
| 16:56:19 | bauzas | sean-k-mooneythe default should what ? | |
| 16:56:37 | sean-k-mooney | bauzas: empty | |
| 16:56:51 | bauzas | empty meaning ? | |
| 16:56:51 | sean-k-mooney | bauzas: placment should not prefer nova over other services | |
| 16:57:20 | sean-k-mooney | meaning unowned | |
| 16:57:29 | bauzas | this is wrong | |
| 16:57:33 | sean-k-mooney | bascaly we ould need a reshape | |
| 16:57:35 | bauzas | any RP is owned by something | |
| 16:57:42 | sean-k-mooney | wel just update | |
| 16:57:48 | sean-k-mooney | to populated it for nova RPs | |
| 16:58:11 | sean-k-mooney | gmann: thanks ill feed that back downstream unless you commented on ghitbub already? | |
| 16:58:33 | bauzas | sean-k-mooney: I need to hardstop but let's continue the convo with gibi later tomorrow | |
| 16:58:34 | gmann | sean-k-mooney: doing on github | |
| 16:58:57 | sean-k-mooney | bauzas: ack o/ | |
| 16:59:03 | sean-k-mooney | gmann: thanks | |
| 16:59:23 | bauzas | sean-k-mooney: I guess I'm afraid by all the implications of tagging our RPs explicitely | |
| 16:59:48 | sean-k-mooney | bauzas: whcih is why we need to discuss in a spec :) | |
| 16:59:53 | bauzas | and I just wanna be sure this is the only possibility | |
| 17:00:21 | bauzas | sean-k-mooney: think about something until tomorrow | |
| 17:01:34 | bauzas | sean-k-mooney: would it be a problem if nova would create RPs with VGPUs in it, while cyborg creates RPs with VGPU in them too, and eventually the allocation would go straight to the nova-compute service holding the allocated RP | |