| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-02-26 | |||
| 11:40:53 | rm_work | like what octavia or live-migrate does | |
| 11:41:08 | sean-k-mooney | the same would be true fo nova careated ports actully since we do that on the compute node | |
| 11:41:11 | rm_work | when we're already locked to a segment due to an existing port | |
| 11:41:22 | sean-k-mooney | ya | |
| 11:41:54 | sean-k-mooney | the only case that breaks is precreated port where you have not set ip_allocaton | |
| 11:42:01 | sean-k-mooney | the only case that breaks is precreated port where you have not set ip_allocaton=defer | |
| 11:43:17 | rm_work | err no, that's exactly one of our cases | |
| 11:43:32 | rm_work | octavia creates a port, it gets an IP... then we need to boot nova VMs in that segment | |
| 11:45:28 | sean-k-mooney | can octavia use ip_allocatio=defer in that case | |
| 11:45:38 | rm_work | no | |
| 11:46:00 | sean-k-mooney | ok then it will work until the cloud start to get full | |
| 11:46:01 | rm_work | the whole point of us creating that port is to pre-allocate an IP up-front | |
| 11:46:03 | sean-k-mooney | or that aggreate | |
| 11:46:09 | rm_work | yes, that's true | |
| 11:46:50 | sean-k-mooney | righ if you preallcoate the ip up front then that determins the segemnt so and with the prefilter we will not consider any other host | |
| 11:47:12 | sean-k-mooney | with all the pros and cons that brings | |
| 11:47:37 | rm_work | yeah that is why this filter is fine for us, we always know the segment we want | |
| 11:47:38 | sean-k-mooney | fundemetally there is no way around that | |
| 11:47:59 | sean-k-mooney | but you dont know if there is enough compute resouces in that segment to run the vm | |
| 11:48:11 | sean-k-mooney | so you will have to handel the no valid host case | |
| 11:48:17 | rm_work | we have no choice | |
| 11:48:21 | sean-k-mooney | and either choose a differnt segment or give up | |
| 11:48:27 | rm_work | usually we are doing this on an object that has existing for a long time | |
| 11:48:47 | rm_work | we have a VIP port, and two VMs that serve HAProxy behind it | |
| 11:48:55 | rm_work | they are in a keepalived pair | |
| 11:49:07 | rm_work | when one fails, we need to boot a new one in the same segment | |
| 11:49:10 | rm_work | there is no choice | |
| 11:49:22 | sean-k-mooney | for routed networks yes | |
| 11:49:31 | sean-k-mooney | although | |
| 11:49:39 | sean-k-mooney | only the vip needs to stay the same | |
| 11:49:46 | sean-k-mooney | sorry | |
| 11:49:59 | sean-k-mooney | octavia the vip is assigned to the vm you are booting | |
| 11:50:48 | sean-k-mooney | rm_work: the vip is the public loadblancer ip right? | |
| 11:51:06 | sean-k-mooney | the one that applciation will use to connect to the services that are behind it | |
| 11:51:38 | rm_work | yes | |
| 11:51:59 | rm_work | the VIP is static and is known to the user | |
| 11:52:07 | sean-k-mooney | yep | |
| 11:52:14 | rm_work | and will always persist (it is unbound, and uses allowed-address-pairs) | |
| 11:52:22 | sean-k-mooney | ya | |
| 11:53:29 | sean-k-mooney | so even in routed network you could use a /32 route to make that available rigth even if the ha proxy vm ip is not in the same subnet range | |
| 11:53:30 | rm_work | so, i'm aware we will have problems if the aggregate fills up, but we have no choice :( | |
| 11:53:50 | rm_work | that'd be possible if we had a different network architecture, maybe | |
| 11:53:56 | rm_work | as it is, we cannot | |
| 11:54:11 | sean-k-mooney | well bgp would be able to supprot that | |
| 11:54:16 | sean-k-mooney | but ya ok | |
| 11:54:35 | mlycka | Hello. Is there a specific person responsible for adding new version templates and folders to nova-spec? | |
| 11:54:52 | sean-k-mooney | not really we have scripts that do it | |
| 11:55:20 | mlycka | Right right...when are you likely to run them for V? | |
| 11:55:39 | sean-k-mooney | mlycka: but you did remind me i need to go update my patch to move implemented specs | |
| 11:55:47 | sean-k-mooney | mlycka: usually not until after m3 | |
| 11:55:58 | rm_work | yeah I would kill for working BGP :D | |
| 11:55:59 | mlycka | m3? | |
| 11:56:26 | sean-k-mooney | milestone 3 so i think start of april | |
| 11:56:49 | mlycka | Crud | |
| 11:57:18 | sean-k-mooney | we just past the spec freeze for U so we usally dont want to start reviewing new spec right away however you could put up a review against the backlog folder and just update it when its created | |
| 11:57:51 | mlycka | I need to move a blueprint proposal to V from U, 'cause I managed to miss my window by being busy and I need to restore it | |
| 11:58:26 | sean-k-mooney | ah ok well let me check something | |
| 11:58:42 | mlycka | Sure, thanks. | |
| 12:00:17 | sean-k-mooney | ok we dont have a script for the new folder although code is welcome. anyone can do it so you have two options. 1 propose a patch that creates the folder and copys the ussuirt template and renames it. 2 restore your spec and propose it to the backlog | |
| 12:00:48 | sean-k-mooney | if you go with option 1 just rebase your current spec on top | |
| 12:01:04 | sean-k-mooney | it just proably wont get looked at for a little bit | |
| 12:01:51 | mlycka | Yeah, that's fair enough. Is the template going to be the same for Victoria then? | |
| 12:03:40 | mlycka | Also, do I need to file a bug for that or is there an existing one or do I just propose a patch without a bug? | |
| 12:07:46 | openstackgerrit | Merged openstack/nova master: trivial: Remove FakeScheduler https://review.opendev.org/707224 | |
| 12:07:53 | openstackgerrit | Merged openstack/nova master: conf: Deprecate '[scheduler] driver' https://review.opendev.org/707225 | |
| 12:08:01 | openstackgerrit | Merged openstack/nova master: docs: Improve documentation on writing custom scheduler filters https://review.opendev.org/707226 | |
| 12:08:13 | openstackgerrit | Merged openstack/nova master: trivial: Use recognized extra specs in tests https://review.opendev.org/708435 | |
| 12:08:36 | gibi | sean-k-mooney, rm_work: mail is up on the ML http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012846.html | |
| 12:10:24 | rm_work | ok | |
| 12:10:52 | rm_work | so basically i guess i just keep running this filter and my patches as-is downstream, and try to participate in getting this moving forward upstream so eventually we can drop it :D | |
| 12:11:38 | gibi | rm_work: good strategy | |
| 12:16:40 | sean-k-mooney | gibi: thanks ill take a look at it in a while | |
| 12:18:11 | sean-k-mooney | gibi: before that however im going to try and go through https://review.opendev.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/provider-config-file today and adress some/all of your comments | |
| 12:18:22 | sean-k-mooney | assuming they are not really hard | |
| 12:20:39 | gibi | sean-k-mooney: cool. ping me if you need some clarification. honestly I lost the context on that so I dont remember if my comments were hard or not | |
| 12:21:58 | sean-k-mooney | well i have not looked at the code before so you know better then i do :) but so far they look ok. i think dustinc is not really around much at the momenet/forseeable future so im going to try and help move it along a bit | |
| 12:24:23 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/rocky: Avoid circular reference during serialization https://review.opendev.org/709798 | |
| 12:24:59 | gibi | this is a stable only bugfix so it need some non stable core to look at ^^ | |
| 12:25:13 | gibi | bauzas, stephenfin: ^^ | |
| 12:29:03 | sean-k-mooney | why is jsonutils.dumps being used instead of to_primitive? | |
| 12:29:55 | openstackgerrit | Marek Lyčka proposed openstack/nova-specs master: Adds spec infrastructure for Victoria https://review.opendev.org/710023 | |
| 12:30:36 | sean-k-mooney | oh i guess you are trying to convert it to a dictionay | |
| 12:30:56 | sean-k-mooney | by round tripping the data through json | |
| 12:31:01 | gibi | sean-k-mooney: the legacy_spec is a dict form of a RequestSpec object | |
| 12:31:12 | sean-k-mooney | ya so we cant just use obj_to_primitive | |
| 12:31:15 | gibi | but it is a special dict form | |
| 12:31:22 | sean-k-mooney | becasue that wont give us what we want | |
| 12:32:09 | gibi | it would give us a different dict | |
| 12:32:44 | sean-k-mooney | yep so it wont work here | |
| 12:32:59 | sean-k-mooney | why do we need to use legacy requests specs in this case? | |
| 12:33:38 | sean-k-mooney | i taught we had already moved to ovo form in rocky but i guess not fully | |
| 12:36:32 | sean-k-mooney | oh its becasue of the workaround for bug #1529084 | |
| 12:36:32 | openstack | bug 1529084 in oslo.messaging "RPC fake driver should accept datetime items for data" [Undecided,Fix released] https://launchpad.net/bugs/1529084 - Assigned to Balazs Gibizer (balazs-gibizer) | |
| 12:37:19 | sean-k-mooney | that intoduced the seriasiation and deserialisation but missed the fact it wont recurse to the nested ovos which is what you are fixing | |
| 12:39:03 | sean-k-mooney | another fix would be to fix the rpc fake dirver to accept datimes in dicst and remvoe the jsonutils calls entirely | |
| 12:39:47 | gibi | sean-k-mooney: since stein we doesn't use the legacy dict but pass around the ovo, but that is an RPC change it is not backportable | |
| 12:40:33 | gibi | sean-k-mooney: but yeah, we we can somehow remove the need of the json.loads(json.dumps(..) stuff that would also help | |
| 12:41:09 | sean-k-mooney | well the comment says its done beacase teh fake rpc driver does not support datetimes in dicts | |
| 12:41:16 | sean-k-mooney | so if we fix that then we could remove it right | |
| 12:41:31 | sean-k-mooney | i assume there was a reason bauzas didnt do that orgininally | |