| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-28 | |||
| 13:25:08 | efried | jaypipes, edleafe: turns out we're not restricting characters on RP names. I just created RPs with names chr(0)-chr(127). Do we have a SQL injection hole, or is sqla providing the necessary escaping security blanket so we don't have to worry about it? | |
| 13:29:22 | jaypipes | efried: sqlalchemy (and in fact the Python DB API) provides the protection against SQL injection attacks. | |
| 13:31:43 | mriedem | dansmith: regarding that [database]/connection setting in nova.conf for the api / controller services, i just noticed that the install guide for the controller tells you to set the [database]/connection to the 'nova' db, rather than nova_cell0 https://docs.openstack.org/nova/latest/install/controller-install-ubuntu.html#install-and-configure-components | |
| 13:32:06 | mriedem | which, if you're not doing 'superconductor' mode, isn't a big deal | |
| 13:32:15 | dansmith | mriedem: probably from pre-cells switchover though | |
| 13:32:25 | dansmith | like we never updated it | |
| 13:33:51 | mriedem | yeah, https://docs.openstack.org/nova/latest/user/cells.html#first-time-setup says, "If your databases are on separate hosts then you should specify --database_connection or make certain that the nova.conf being used has the [database]/connection value pointing to the same user/password/host that will work for the cell0 database." | |
| 13:34:31 | mriedem | with the current install guide instructions, i don't think cell0 db sync ever runs | |
| 13:35:27 | dansmith | ack, we should probably audit all that stuff in modern times | |
| 13:37:29 | mriedem | oh nvm, db sync defaults to find the cell0 mapping and sync it | |
| 13:37:51 | olivenwk | Hello, I try to add aliases in nova. Why alias ={ "vendor_id":"1002", "product_id":"6938", "device_type":"type-PCI", "name":"R9-380X" } followed by alias ={ "vendor_id":"10de", "product_id":"1b06", "device_type":"type-PCI", "name":"GTX1080Ti" } works and not alias=[{ "vendor_id":"1002", "product_id":"6938", "device_type":"type-PCI", "name":"R9-380X" },{ "vendor_id":"10de", "product_id":"1b06", "device_type":"type-PCI", "name":"GTX1080Ti" }] | |
| 13:38:36 | olivenwk | but if I write my alias in the array fashion ONLY on the compute node it works | |
| 13:38:55 | olivenwk | seems to be controller node related, any ideas? | |
| 13:45:42 | openstackgerrit | Konstantinos Samaras-Tsakiris proposed openstack/nova master: Add `hide_hypervisor_id` flavor extra_spec https://review.openstack.org/555861 | |
| 13:52:17 | openstackgerrit | Matt Riedemann proposed openstack/nova-specs master: Allow abort live migrations in queued status https://review.openstack.org/536722 | |
| 13:52:56 | gibi | jaypipes: I left two questions in the nested a_c spec https://review.openstack.org/#/c/556873/ | |
| 13:55:24 | mriedem | eric-young: https://review.openstack.org/#/c/554679/ is now released in os-brick 2.4.0, you should push a change to the openstack/requirements repo to bump the minimum required version to 2.4.0 in global-requirements | |
| 13:55:38 | mriedem | and then make https://review.openstack.org/#/c/554679/ depend on it | |
| 13:58:47 | dansmith | hmm, someone removed the placement-req series from the runways etherpad queue | |
| 13:59:07 | dansmith | oh, nevermind, it got moved down apparently | |
| 14:00:03 | mriedem | heh, the glory of etherpads | |
| 14:00:11 | mriedem | pre-emptible runways | |
| 14:00:47 | mriedem | so uh, do we wait for melwitt to move things from the queue into the actual slots? | |
| 14:01:20 | dansmith | yeah I went to look because I expected she moved things into the slots after yesterday | |
| 14:01:36 | dansmith | we have a meeting at 8 so she should be around then-ish I think | |
| 14:01:57 | mriedem | there is a procedure on L24 so i guess anyone could just follow that | |
| 14:02:15 | openstackgerrit | Eric Fried proposed openstack/nova master: Slugification utilities for placement names https://review.openstack.org/556628 | |
| 14:02:30 | bauzas | do we really need slots for priority series ? | |
| 14:02:40 | mriedem | efried: please confirm your availability for the update_provider_tree bp for runways slot numero uno for the next dos weeks | |
| 14:02:48 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Allow scheduling only to enabled cells (Filter Scheduler) https://review.openstack.org/550527 | |
| 14:02:49 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Add --enable and --disable options to nova-manage update_cell https://review.openstack.org/555416 | |
| 14:02:50 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Update the cells FAQs and scheduler maintenance docs. https://review.openstack.org/556459 | |
| 14:02:55 | efried | cdent: edleafe: Any idea how I can finagle my string so that json encoding doesn't bounce with UnicodeDecodeError for characters >127? | |
| 14:03:01 | efried | mriedem: Confirmed. | |
| 14:03:02 | bauzas | I mean, I can review those priority series without needing a slot | |
| 14:03:07 | efried | mriedem: Confirmado? | |
| 14:03:53 | bauzas | because I know about the implementations, and I don't need to understand it | |
| 14:04:04 | cdent | efried: I'm non in-conext. Have you got code to point at? | |
| 14:04:13 | mriedem | ANNOUNCEMENT: https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/update-provider-tree is now in runway slow #1 | |
| 14:04:25 | efried | mriedem: Nice freudian slip there. | |
| 14:04:34 | mriedem | ha | |
| 14:04:39 | mriedem | oh well, it's not recorded | |
| 14:04:45 | mriedem | i mean, it is | |
| 14:04:48 | mriedem | but not via bot | |
| 14:04:58 | mriedem | is kashyap going to write a runway bot? | |
| 14:05:10 | mriedem | that would actually be kind of nice...to post to a wiki rather than etherpad | |
| 14:05:21 | dansmith | we could tweak the channel topic | |
| 14:05:37 | dansmith | I dunno that release schedule is that necessary as it's easily google-able | |
| 14:05:44 | efried | cdent: In writing https://review.openstack.org/#/c/556628/ I was brought to realize we're not restricting characters for RP names. I tried chr(0)-chr(127) and they all worked fine. But when I start trying chr(128)-chr(256) I get UnicodeDecodeError from the ksa session json encoder. | |
| 14:06:03 | cdent | 128-256 is not valid unicode | |
| 14:06:19 | mriedem | dansmith: or just link https://etherpad.openstack.org/p/rocky-nova-priorities-tracking in the channel topic and link the runways etherpad from the top of that | |
| 14:06:31 | mriedem | or a gerrit dashboard link, | |
| 14:06:42 | mriedem | where mel controls the starred patches per runway that go into said dashboard | |
| 14:06:43 | mriedem | or something | |
| 14:07:03 | dansmith | well, it'd just be nice to see names as reminders instead of the same url week to week, is what I meant | |
| 14:07:23 | dansmith | (hence the irc announcement part of that procedure) | |
| 14:07:36 | mriedem | so like, current runways: update-provider-tree, placement-req, foobars | |
| 14:07:41 | dansmith | cha | |
| 14:07:53 | openstackgerrit | Elod Illes proposed openstack/nova stable/ocata: Set error state after failed evacuation https://review.openstack.org/518733 | |
| 14:07:53 | openstack | bug 1713783 in OpenStack Compute (nova) ocata "After failed evacuation the recovered source compute tries to delete the instance" [High,In progress] https://launchpad.net/bugs/1713783 - Assigned to Balazs Gibizer (balazs-gibizer) | |
| 14:07:53 | openstackgerrit | Elod Illes proposed openstack/nova stable/ocata: Functional test for regression bug #1713783 https://review.openstack.org/505160 | |
| 14:08:07 | mriedem | wfm | |
| 14:08:55 | efried | cdent: So JSON strings have to be unicode I take it. | |
| 14:08:57 | mriedem | so i guess next in line is privsep | |
| 14:09:22 | cdent | efried: it depends on which layer you're at, over the wire JSON is supposed to always be utf-8 encode unicode | |
| 14:09:52 | mriedem | mikal: please confirm your availability for https://blueprints.launchpad.net/nova/+spec/hurrah-for-privsep-again reviews for the next 2 weeks | |
| 14:10:25 | mriedem | likely not to hear about that for several hours | |
| 14:10:27 | dansmith | mriedem: heh, that answer won't come for 10 hours | |
| 14:10:28 | dansmith | yeah | |
| 14:10:45 | cdent | efried: once you see the results in python, it should be unicode strings (or numbers) | |
| 14:10:53 | dansmith | since the queue is fresh I tend to think we can assume in this case, but... | |
| 14:11:08 | mriedem | dansmith: please confirm your availability for https://blueprints.launchpad.net/nova/+spec/placement-req-filter reviews for the next 2 weeks | |
| 14:11:42 | dansmith | mriedem: I will be available with the exception of this coming friday which is a company holiday, but I don't anticipate that being significant | |
| 14:11:52 | mriedem | EASTER! | |
| 14:11:54 | efried | dansmith, mriedem: FWIW mikal has been responsive to those patches within a couple days up to this point. And with I think one or two exceptions, they've all been scrubbed to multiple-+1-ness at this point. | |
| 14:12:04 | dansmith | efried: aye | |
| 14:12:11 | mriedem | CONFIRMED | |
| 14:12:12 | edleafe | Heh - "slow" [t2LP0] | |
| 14:12:35 | edleafe | Lemme fix that [t 2LP0] | |
| 14:12:36 | purplerbot | ||
| 14:12:46 | dansmith | heh | |
| 14:13:22 | efried | dansmith: I happened to notice that mikal (whether on purpose or not) changed the commit message tags from 'hurrah-for-privsep-again' to 'hurrah-for-privsep' in the last spin. | |
| 14:13:34 | efried | I didn't go and check what the bp is actually called. | |
| 14:13:42 | dansmith | efried: that's likely git-review | |
| 14:13:52 | efried | it was in the commit message, not the topic | |
| 14:15:02 | dansmith | oh | |
| 14:18:10 | bhagyashris | mriedem, alex_xu_: Hi, Thanks for review! Addressed review comments on spec https://review.openstack.org/#/c/511825/ request to re-visit | |
| 14:18:59 | mriedem | ok | |
| 14:21:57 | kashyap | 20 is the lucky number, if anyone looking for a trivial feature to +2 (it already has +2 from Alex): https://review.openstack.org/#/c/534384/ | |
| 14:23:14 | kashyap | Zuul has to bless & kiss it, though | |
| 14:23:34 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/pike: Remove usage of kwarg retry_on_request in API https://review.openstack.org/557348 | |
| 14:25:06 | bauzas | efried: I don't know if it's you that wrote https://etherpad.openstack.org/p/nova-runways-rocky L39 but I think virt driver changes using nested resources shoukdn't be using that slot | |
| 14:25:22 | mriedem | the placement aggregates api confuses me every time i look at it | |
| 14:25:24 | efried | bauzas: It was, and I agree they shouldn't. | |
| 14:25:38 | efried | mriedem: How so? Maybe we can fix the docs. | |
| 14:25:39 | bauzas | efried: because those need specific domain expertise, and I think it's a separate runway slot then | |
| 14:26:05 | efried | bauzas: They're also a separate piece of work, *using* the upt bp, not part of its implementation. | |
| 14:26:06 | mriedem | efried: i think it's just because there is no POST for aggregates | |
| 14:26:12 | mriedem | but i know that's intentional | |
| 14:26:22 | efried | mriedem: Yeah, they get created (but not deleted!) dynamically. | |