Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-03-30
14:15:26 jaypipes ralonsoh: ah, gotcha.
14:15:52 bauzas sdague: oh man, did I say I love your blogpost ? https://dague.net/2017/03/30/in-praise-of-systemd/
14:16:03 jaypipes ralonsoh: you might want to ping johnthetubaguy, dansmith, mriedem and bauzas on that spec since they are nova-specs cores.
14:16:18 bauzas sdague: *that* is the exact feature I wanted without needing to make my own silly script
14:16:27 ralonsoh jaypipes: thanks for the tip!
14:16:30 jaypipes heh, I bet mordred loves sdague's blog post ;)
14:18:15 jianghuaw__ jaypipes, could you have a look at this proposal on vgpu when you have time? https://review.openstack.org/#/c/450122/
14:18:31 mordred jaypipes, sdague: honestly, it's a good post and is literally the first explanation of an actual feature that is provided that I cannot immediately naysay with how we've always been able to do that - and also easier
14:18:32 jaypipes jianghuaw__: yep! it's on my docket for the day.
14:19:01 dansmith mordred: *gasp*
14:19:17 dansmith mordred: you just abandoned your whole belief system
14:19:23 mriedem Kevin_Zheng: ok i might work on a fix then if i find the time on my flight today,
14:19:27 jianghuaw__ jaypipes, thanks.
14:19:29 mriedem Kevin_Zheng: i think we just want to use _local_delete
14:19:38 ralonsoh johnthetubaguy, dansmith, mriedem, bauzas: hi. I have a spec to be reviewed (#link https://review.openstack.org/#/c/435954/). I you have time, can you have a look?
14:19:39 mriedem that will handle port deallocation and volume cleanup
14:19:42 mordred jaypipes, sdague: I do very much wish the journald format wasn't binary requiring a tool to read, that it wasn't prone to corruption on server crash and that it wasn't tied together with a DAG-based event driven init system that makes operating things harder
14:19:50 mordred dansmith: ^^ there you go
14:20:05 dansmith heh
14:20:07 ralonsoh johnthetubaguy, dansmith, mriedem, bauzas: I think it's almost finished
14:20:30 bauzas mordred: well, while you have valid concerns, I think that's still a good move worth it :)
14:20:39 mordred because let's be frank - just making a thing that's similar to syslog but structured like what sdague wrote up is awesome - but tying all of linux together with a giant monolithic stack that touches everything is not required to get better logging
14:20:53 Kevin_Zheng mriedem: thanks, I will check it :) just don't know the best way to implement it
14:21:02 dansmith my belief system is: mordred is right about 99% of things, I'm willing to let him be wrong about systemd
14:21:11 bauzas hah
14:21:19 mordred dansmith: that's a good belief system - it saves you the most amount of pain :)
14:21:23 dansmith heh
14:21:27 sdague woah, I got event faint praise from mordred around systemd :)
14:21:34 jaypipes :)
14:21:37 mordred sdague: I know right?
14:21:37 sdague I consider that a win
14:21:50 dansmith sdague: it's more than I hope to get from mordred on any topic
14:21:55 mordred sdague: you know - using systemd as a user-space non-pid-1 application launcher is a fine idea
14:22:03 bauzas ralonsoh: on ity
14:22:12 ralonsoh bauzas: thanks!
14:22:12 mordred and I wish that was how we had it rolled out to linux
14:22:30 stackempty Hi guys .. does anyone know if this https://specs.openstack.org/openstack/nova-specs/specs/newton/approved/flavor-root-disk-none.html has been implemented .. or if there's some other way to achieve it.
14:26:55 sdague bauzas: also, just to be clear, all the bits to do what I said in that blog post have not yet landed
14:27:07 sdague and are off by default right now anyway
14:27:24 bauzas sdague: I know, I CC'd myself on the oslo.log change
14:27:28 sdague bauzas: cool
14:28:08 migi dansmith: quick question about subsequent runs of nova-manage archive_deleted_rows, the output is as follows, to ensure CLI works as expected should I call it until "nothing was archived" is reached or first run is enough and then see if some of the rows were archived ? http://paste.openstack.org/show/604862/
14:28:24 openstackgerrit Gergely Csatari proposed openstack/nova master: api-ref: adminPass is optional in Create Server https://review.openstack.org/451497
14:29:34 dansmith migi: operators may not want to run it until they get to that point, because that will purge everything out of the db, but for testing, we should confirm that you _can_ run it until you get to that point, yeah
14:29:48 migi dansmith: ok
14:35:47 bauzas jaypipes: so you're cool with us adding more capabilities into the blob that's stored in the DB for PCI devices ?
14:36:02 bauzas jaypipes: doesn't that impact your work on cleaning that up ?
14:37:12 jaypipes bauzas: see my comments on that review from earlier. basically, I'm fine adding those NIC features as a temporary effort until we get traits done (and have the flavor able to communicate required and preferred traits)
14:37:25 jaypipes bauzas: more of a compromise than anything else, really.
14:37:31 ralonsoh bauzas: once os-traits and resource providers is working, we will migrate it
14:37:49 bauzas jaypipes: okay, just wanted to make sure you balanced that
14:37:57 bauzas - which I was pretty sure of tho - :)
14:38:55 bauzas jaypipes: FWIW, just reviewing https://review.openstack.org/#/c/448283/ as a side one :)
14:39:10 sfinucan jaypipes, bauzas, ralonsoh: ^ if that point's not in the spec, it should be
14:39:29 sfinucan the "this is a temporary solution until placement can take over" bit, that is
14:39:42 bauzas sfinucan: that's just a gerrit comment for the moment
14:40:10 bauzas sfinucan: I'm not violently opposed on merging that that way
14:40:23 bauzas sfinucan: given the gerrit comment is visible
14:40:41 ralonsoh sfinucan: https://review.openstack.org/#/c/435954/11/specs/pike/approved/enable-sriov-nic-features.rst, L222
14:40:41 bauzas only if we were having a new PS, I'd ask for that being in the spec
14:40:43 sfinucan bauzas: I should hope you're not "violently opposed" to anything in OpenStack :D
14:41:13 sfinucan ralonsoh: Ah, good stuff :)
14:41:26 bauzas sfinucan: I'm violently opposed to bureaucracy, which is hopefully not a trait of OpenStack :)
14:41:34 ralonsoh sfinucan: :P
14:42:21 bauzas ralonsoh: I'm missing how you express the NIC offload capabilities when booting in the spec
14:42:29 bauzas ralonsoh: same old story I guess with flavors ?
14:42:48 bauzas nevermind L127
14:42:56 ralonsoh bauzas: flavors is now out of spec
14:43:05 ralonsoh bauzas: yes, using a neutron port
14:43:41 sfinucan edleafe: Aye, death to the fanatics
14:44:11 bauzas edleafe: I can be extremist for food
14:44:37 sfinucan jaypipes: If you're looking at nova-specs, this is a little improvement on an existing spec https://review.openstack.org/#/c/447401/
14:44:50 sfinucan *one of your existing specs
14:48:20 bauzas ralonsoh: just to understand correctly, we need to define a neutron port (with mentioning its binding profile) before doing nova boot with the flavor expressing the asked capabilities ?
14:54:33 jaypipes sfinucan: reviewed.
14:58:19 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: SSP emphemeral disk support https://review.openstack.org/443189
14:59:24 openstackgerrit Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for GET RP inventories https://review.openstack.org/449765
14:59:25 openstackgerrit Andrey Volkov proposed openstack/nova master: [placement] Add api-ref for POST RP inventories https://review.openstack.org/451855
15:02:14 openstackgerrit Lee Yarwood proposed openstack/nova master: encryptors: Switch to os-brick encryptor classes https://review.openstack.org/391597
15:04:11 openstackgerrit Eric Fried proposed openstack/nova master: Glance download: only fsync files https://review.openstack.org/451094
15:04:47 efried mdbooth Good call on those square brackets. ^^
15:06:46 openstackgerrit Lee Yarwood proposed openstack/nova master: compute: Detach volumes on _rebuild_default_impl failure https://review.openstack.org/442105
15:08:37 ralonsoh bauzas: sorry, I was afk. That's almost correct: 1) creation of neutron port, 2) define pci filter (with these new parameters) and 3) boot the machine
15:08:52 ralonsoh bauzas: you don't need to modify the flavor
15:11:25 ralonsoh bauzas: when I say "define pci filter" I mean to enable it
15:15:46 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: config drive https://review.openstack.org/409404
15:15:47 mriedem apparently it doesn't take much time to get to or through SJC and here i sit with all of this idle time
15:16:46 openstackgerrit Stephen Finucane proposed openstack/nova-specs master: PCI NUMA Policies https://review.openstack.org/361140
15:18:10 sdague mriedem: do you want to adopt any of this language to your quota doc - https://review.openstack.org/#/c/450244/ ? I had mostly meant it to be an example, but it looks like the other doc thing merged
15:18:50 bauzas ralonsoh: so, I +W'd it but I have a question https://review.openstack.org/#/c/435954/11/specs/pike/approved/enable-sriov-nic-features.rst@236
15:19:23 jaypipes ralonsoh: spec is merged. you can un-WIP your patch series :)
15:19:39 ralonsoh bauzas: yes, I'm on it. I know this is the trickiest point
15:19:46 ralonsoh jaypipes: thanks!
15:20:17 openstackgerrit Anusha Unnam proposed openstack/nova master: Add keypairs field to InstancePayload https://review.openstack.org/419730
15:20:21 mriedem ralonsoh: what is the current state of intel 3rd party nfv ci since wznoinsk left?
15:20:33 mriedem before we talk about adding more new nfv things and new ci
15:20:52 mriedem sdague: maybe? i was going to try and spend some time reading your limits spec
15:21:04 ralonsoh mriedem: this is still in progress. Now I can't tell you (I don't know) which is the current status
15:21:12 sdague mriedem: yeh, the limits one is more important
15:21:28 vladikr jaypipes, bauzas, hey could you please take a look at the xen vgpu spec, when you have a moment. https://review.openstack.org/#/c/450122 - I'm trying to convince the author to make the spec more generic, so we could implement the libvirt side of the story under the same spec

Earlier   Later