| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-12 | |||
| 14:00:16 | gibi | stephenfin: scheduler meeting starts now, I will ping you when it is over | |
| 14:00:38 | dansmith | kashyap: I have it open, I'll look at it after I dig out from the morning | |
| 14:00:42 | dansmith | johnthetubaguy: you gonna hit it first? | |
| 14:00:56 | edleafe | Scheduler subteam meeting running now in #openstack-meeting-alt | |
| 14:00:59 | kashyap | dansmith: Sure, take your sweet time. I need to rework the unit tests anyway. | |
| 14:01:08 | dansmith | kashyap: it would really help if it was passing unit tests :) | |
| 14:01:08 | kashyap | dansmith: Also, additional context (that I cut down) is here: https://review.openstack.org/#/c/534384/2//COMMIT_MSG | |
| 14:01:09 | dansmith | hah, yeah | |
| 14:01:30 | kashyap | dansmith: Sure, don't open it until I fix the unit tests :-) | |
| 14:02:10 | ShilpaSD | Stephenfin: hi, actually want to inform you that i am working on issue https://github.com/novnc/noVNC/issues/967, and for this updating nova\compute\manager.py for path as suugested on github issue | |
| 14:02:35 | ShilpaSD | so is it okay to chang path in such way? or is there any way around? | |
| 14:03:12 | johnthetubaguy | dansmith: yeah, looking at it now | |
| 14:03:52 | ShilpaSD | https://github.com/openstack/nova/blob/d7c46b279686308f5431410bfccb3e40031e380a/nova/compute/manager.py#L5058 | |
| 14:04:07 | kashyap | johnthetubaguy: Many thanks. ACK / NACK on the core change in driver.py would be useful. (I'm still looking at what tests got broken & a new unit test.) | |
| 14:05:51 | stephenfin | ShilpaSD: Yeah, I tried to reproduce that and couldn't. I only needed to make one change to the configuration. See my notes here https://review.openstack.org/#/c/483994/ | |
| 14:06:02 | stephenfin | ShilpaSD: It's the comments from March 6th | |
| 14:06:44 | kashyap | johnthetubaguy: But you're right that I should've retained the URL to the perf degradation analysis. Feel free to call that out, I'm fixing it locally | |
| 14:06:54 | stephenfin | ShilpaSD: Perhaps I missed something though. Could you provide compare my steps to yours to see if this is the case? | |
| 14:07:14 | stephenfin | johnthetubaguy: Sure, that would be great. I'll go through and fix all those up now :) | |
| 14:08:22 | openstackgerrit | Merged openstack/nova master: XenAPI: XCP2.1+ Swallow VDI_NOT_IN_MAP Exception https://review.openstack.org/538415 | |
| 14:08:30 | openstackgerrit | Merged openstack/nova master: Make nova build reproducible https://review.openstack.org/551269 | |
| 14:09:41 | ShilpaSD | stephenfin: hi, i have gone through https://review.openstack.org/#/c/483994/9/nova/conf/vnc.py | |
| 14:09:51 | ShilpaSD | we need to use vnc_lite.html | |
| 14:10:06 | ShilpaSD | i have verified using vnc.html | |
| 14:10:15 | ShilpaSD | and there we have issue | |
| 14:10:34 | stephenfin | ShilpaSD: Oh, ok. So if you use 'vnc_lite.html', it works as expected? | |
| 14:11:13 | stephenfin | ShilpaSD: Does 'vnc.html' work with noVNC < 1.0.0? If so, we might have a migration path to avoid people breaking when they upgrade to noVNC 1.0 | |
| 14:11:15 | ShilpaSD | not tested with vnc_lite.html yet | |
| 14:12:33 | ShilpaSD | i have gone through your changes, and from there i thought you verified for vnc_lite.html, wait will check with vnc_lite and get back to you. | |
| 14:13:56 | stephenfin | ShilpaSD: So they renamed 'vnc_auto.html' to 'vnc_lite.html' in 1.0. Other than the rename, it works exactly the same as before | |
| 14:14:21 | stephenfin | ShilpaSD: However, if 'vnc.html' works for both 0.6 and 1.0, we might need your fix and then we can change the default from 'vnc_auto.html' to 'vnc.html' | |
| 14:14:46 | stephenfin | That means people could use either noVNC 0.6 or 1.0 with no other config changes, which would make upgrading much easier | |
| 14:14:51 | stephenfin | Let me know what you find :) | |
| 14:15:11 | ShilpaSD | stephenfin: okay, will test and get back to you | |
| 14:27:36 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove deprecated 'multi_instance_display_name_template' opt https://review.openstack.org/499612 | |
| 14:27:54 | sean-k-mooney | stephenfin: is novnc running under a webserver? if so adding a redirect from vnc_auto.html to vnc_lite.html should be pretty simple | |
| 14:28:19 | stephenfin | sean-k-mooney: That...is not a bad idea | |
| 14:28:35 | stephenfin | Lemme have a look at how DevStack is deploying it | |
| 14:28:40 | arvindn05 | mriedem: https://review.openstack.org/#/c/541507/ - can you take a look at the latest patchset? I think it should address your concerns | |
| 14:40:30 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove deprecated 'allow_instance_snapshots' opt https://review.openstack.org/499621 | |
| 14:41:16 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove 'db_driver' config opt https://review.openstack.org/508487 | |
| 14:41:35 | cdent | stephenfin++++++++++ | |
| 14:42:02 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Fix indentation of database options https://review.openstack.org/443097 | |
| 14:42:38 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Resolve TODOs in 'database' https://review.openstack.org/393695 | |
| 14:42:46 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Resolve TODOs in 'database' https://review.openstack.org/393695 | |
| 14:42:55 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove 'db_driver' config opt https://review.openstack.org/508487 | |
| 14:43:17 | bauzas | heh, Credit Impot Recherche, rather | |
| 14:44:08 | stephenfin | johnthetubaguy: I think I've most of those rebased and the out-of-date blueprint line removed. If you fancy pushing them through, that would be appreciated :) | |
| 14:44:40 | johnthetubaguy | stephenfin: on a call, but will try hit those, do they still have some shared topic? | |
| 14:44:59 | stephenfin | johnthetubaguy: Sure. I kept the topic the same https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/centralize-config-options-queens | |
| 14:45:04 | edmondsw | stephenfin I had to manually rebase after you +Wd... can you take another look at https://review.openstack.org/#/c/539314/ ? | |
| 14:47:19 | stephenfin | edmondsw: Sure | |
| 14:47:27 | efried | melwitt: The "runways" thing - is that going to be tracked on https://etherpad.openstack.org/p/rocky-nova-priorities-tracking or is there a separate etherpad? | |
| 14:47:33 | edmondsw | stephenfin ty sir | |
| 14:48:42 | ShilpaSD | stephenfin: Hi, tested vnc console access using 'vnc_lite.html' for version 1.0, its working without any changes at nova-compute services, and able to access console | |
| 14:49:07 | stephenfin | ShilpaSD: Excellent. So there's nothing more needed, if we decide to use that? | |
| 14:49:11 | ShilpaSD | but if configured 'vnc.html' for version 1.0, then require changes at nova-compute service | |
| 14:49:21 | stephenfin | What about version 0.6? | |
| 14:49:52 | ShilpaSD | 0.6 not yet checked, will do that in next 5 min and get back to you | |
| 14:49:56 | stephenfin | (y) | |
| 14:50:33 | ShilpaSD | should i use v v0.6.1 or v0.6.2? | |
| 14:53:58 | ShilpaSD | stephenfin: actually earlier we had 'vnc_auto.html', and that removed in currunt vnc 1.0 release, so we have to make configuration chnages | |
| 14:55:02 | stephenfin | ShilpaSD: Yup, that what's I found. What I'm suggesting is that maybe 'vnc.html' works in both 0.6 and 1.0. If so, we might start using that instead of 'vnc_auto' or 'vnc_lite' | |
| 14:55:26 | stephenfin | Also, use 0.6.2 I guess. Either would be OK though | |
| 14:56:29 | ShilpaSD | okay, will check on 0.6.0 with 'vn.html' and get back to you with outcome | |
| 14:56:43 | ShilpaSD | 'vnc.html'* | |
| 15:01:07 | jaypipes | bauzas, cdent: can you guys give me 45 minutes? | |
| 15:01:09 | bauzas | efried: I don't see why it could be "madness" to accept operator's defined traits | |
| 15:01:31 | cdent | jaypipes: you mean before continuing this topic? I can, sure | |
| 15:01:34 | bauzas | the munging logic should be very explicit | |
| 15:01:35 | jaypipes | yes | |
| 15:01:40 | bauzas | jaypipes: yup, I can hold | |
| 15:03:39 | gibi | jaypipes, stephenfin: so the scheduler meeting is over. I have about two hours left from my workday to jump on a hangouts about the numa aware vswitches | |
| 15:04:15 | stephenfin | gibi: I think jaypipes is busy for a bit so maybe once he's free? I've no other meeting scheduled for the day | |
| 15:04:34 | gibi | stephenfin: OK. I just noted down my boundaries :) | |
| 15:05:22 | stephenfin | gibi: Sure, we'll fit it in in that time :) | |
| 15:12:20 | ShilpaSD | stephenfin: Hi, for 'vnc.html' with noVnc version 0.6.2 and websockify-0.8.0, it failed to access, error at do_handshake() | |
| 15:12:40 | stephenfin | ShilpaSD: Is that the same issue you saw with 1.0? | |
| 15:13:09 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Standardize '_get_XXX_constraint' functions https://review.openstack.org/385071 | |
| 15:13:54 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Move availability zones opts to a group https://review.openstack.org/462469 | |
| 15:17:41 | ShilpaSD | for 1.0, https://github.com/novnc/noVNC/issues/967 | |
| 15:17:52 | ShilpaSD | actually issue for noVNC=v1.0.0-testing.2 | |
| 15:19:22 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove pagesize from __init__ of InstanceNUMATopology https://review.openstack.org/485553 | |
| 15:19:22 | ShilpaSD | for 1.0, token is not available as query parameter as well as it is not set in cookie. | |
| 15:19:22 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove related pinning from __init__ of InstanceNUMATopology https://review.openstack.org/485554 | |
| 15:19:23 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove cpuset_reserved from __init__ of InstanceNUMATopology https://review.openstack.org/466030 | |
| 15:20:06 | ShilpaSD | and if we configure 'vnc.html' | |
| 15:22:11 | stephenfin | ShilpaSD: So they're different issues? | |
| 15:26:24 | ShilpaSD | stephenfin: No, its same. token is not getting as query parameter as well as it is not set in cookie. | |
| 15:26:43 | ShilpaSD | for vnc.html configuration | |
| 15:27:34 | ShilpaSD | stephenfinr: please firstly confirm me that are we going to use vnc_lite.html or vnc_html for version 1.0? | |
| 15:28:10 | ShilpaSD | token unavailable issue is with vnc.html only | |
| 15:33:01 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Don't use single backticks in man pages https://review.openstack.org/540887 | |
| 15:33:02 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Start using openstackdoctheme's extlink extension https://review.openstack.org/540888 | |
| 15:33:49 | stephenfin | ShilpaSD: So, this is my idea. We fix the issue you're seeing and change the default to 'vnc.html'. That means people can use the same configuration file with both 0.6 and 1.0 | |
| 15:34:36 | stephenfin | I don't know if that's a good idea but, in my mind, this would make for the easiest upgrade path for operators | |
| 15:36:26 | sean-k-mooney | stephenfin: since upgrades of openstack servers are supposed to be doable without any config changes from n to n+1 i think it makes sense | |
| 15:37:21 | stephenfin | sean-k-mooney: Right. We could conceivably backport that fix too to make things even easier for the N to N+M cases | |
| 15:37:38 | ShilpaSD | stephenfin: but earlier its 'vnc_auto.html' for 0.6, so configuration file changes are require with my fixes too | |
| 15:38:26 | sean-k-mooney | stephenfin: well for n to n+2 it is expect that config change could be required but that should be achive via an itermediate change during n+1 | |