Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-01-17
12:37:50 sean-k-mooney on a slightly different tangent. for https://review.openstack.org/#/c/465160/ we were waiting for the intel nfv ci to run. that shoudl be up before the ptg however the ci-worker vm i was doing the os-vif work in was created to allow me to replicate how the nfv ci would work.
12:38:29 sean-k-mooney if i get time later today ill try and test the latest version uploaded by tetsuro and see if i can upload the logs somewhere
12:39:06 sean-k-mooney my question is is that sufficent or do we want to leave this to rocky when the ci should be back
12:40:23 jaypipes sean-k-mooney: heh, I stumbled across that patch again yesterday closing out old browser tabs and wondered what the latest was on it.
12:41:10 jaypipes sean-k-mooney: I don't consider it to be a high priority, frankly. If that needs to bump to Rocky as we wait for functional ci, I'm very much ok with that
12:42:31 sean-k-mooney jaypipes: ok Tetsuro seams to have a serise of patches that build on it to be a more complete solution so proably better to wait for the whole set instead of just have the first patch
12:43:20 jaypipes sean-k-mooney: ya
12:44:59 sean-k-mooney cool thats one thing i can take of my too do list for today. so jay if you are happy with https://review.openstack.org/#/c/534307/ do you want me to propose a 1.8.1 release to the release repo with that commit
12:47:58 jaypipes sean-k-mooney: +2 from me. I'll let stephenfin +Wallaby it.
12:50:48 sean-k-mooney jaypipes: cool, its parent is not the had of master so it will create a merge commit with a new sha so ill wait for it to be merged the ill update this https://github.com/openstack/releases/blob/master/deliverables/queens/os-vif.yaml
12:51:10 sean-k-mooney should i use version: 1.8.1 or 1.9.0
12:51:27 ameeda jaypipes: can you please check my comments here ? "https://review.openstack.org/#/c/526900/"
12:52:40 jaypipes ameeda: I did. I still do not understand your response on https://review.openstack.org/#/c/526900/19/nova/utils.py.
12:54:16 ameeda I am sorry, I meant by that, the image has key called "value" , so I filter the request to make sure that I will not change another key limitations.
12:55:02 jaypipes ameeda: why does the image have a key called "value"? that doesn't make sense.
12:55:44 ameeda seriously I don't know, but that what I found in the code
12:56:30 sean-k-mooney it forced reboots are awsome... brb
12:57:13 ameeda jaypipes: you can see that here also "https://review.openstack.org/#/c/526900/20/nova/tests/functional/regressions/test_bug_1117923.py" Line 114
12:58:16 jaypipes ameeda: that's only because you are hard-coding the key to "value" on line 73.
12:59:05 ameeda oh sorry ! yes I got that, so I have to remove the key!="value" ?
12:59:36 jaypipes ameeda: yessir
13:00:16 ameeda jaypipes: so the code should truncates all values of keys ?
13:00:34 jaypipes ameeda: even then, though, keep in mind the REST API for compute will limit any input from users to 255 characters. I understand that the image metadata values are not set through the REST API, though
13:00:59 jaypipes ameeda: no, I'm saying you should *not* truncate the values in that utils module function.
13:01:27 ameeda jaypipes: yes that what I would to say.
13:01:44 jaypipes ameeda: rely on the REST API to filter out user-supplied values >255 chars in length and allow non-user-supplied metadata values (like those coming from Glance) to be passed as-is to Nova's database.
13:02:31 ameeda jaypipes: what you guess me to write ? how I can do that ?
13:03:51 jaypipes ameeda: other than the test case change (don't hard-code the key to 'value'), the only change you need to make is in here:
13:03:52 jaypipes https://review.openstack.org/#/c/526900/20/nova/utils.py
13:04:22 jaypipes ameeda: simply remove original line 983 and change line 984 to use "value" instead of "new_value"
13:05:34 ameeda that will be acceptable, so you can give me +2 :D ?
13:05:34 jaypipes ameeda: left notes in the review for you.
13:05:44 ameeda jaypipes: thanks
13:11:09 ameeda jaypipes: its good to name the new key as "new_key" rather than "value" ?
13:12:20 jaypipes ameeda: call the key something realistic that would be set by Glance.
13:13:12 ameeda "new_key_from_glance" for example ?
13:14:21 ameeda jaypipes: its fine ?
13:15:39 jaypipes ameeda: how about something realistic, like "trusted_certificates" and make the value of the metadata item 2K of random bytes.
13:17:02 ameeda I have to do 2k random bytes by code or hard coded ?
13:17:37 openstackgerrit Matthew Booth proposed openstack/nova master: Give volume DriverBlockDevice classes a common prefix https://review.openstack.org/526346
13:17:38 openstackgerrit Matthew Booth proposed openstack/nova master: Expose driver_block_device fields as attributes https://review.openstack.org/528362
13:17:38 openstackgerrit Matthew Booth proposed openstack/nova master: Add DriverLocalImageBlockDevice https://review.openstack.org/526347
13:17:39 openstackgerrit Matthew Booth proposed openstack/nova master: Pass DriverBlockDevice to driver.attach_volume https://review.openstack.org/528363
13:17:39 openstackgerrit Matthew Booth proposed openstack/nova master: Add local_root to block_device_info https://review.openstack.org/529029
13:17:40 openstackgerrit Matthew Booth proposed openstack/nova master: Use real block_device_info data in libvirt tests https://review.openstack.org/527916
13:18:38 openstackgerrit Merged openstack/python-novaclient master: Add missing spaces in `nova list --changes-since` help https://review.openstack.org/533575
13:20:11 ameeda jaypipes : this code is okay for you ? "import random def randomBytes(n): return bytearray(random.getrandbits(8) for i in range(n))"
13:22:21 jaypipes ameeda: sure, that's fine.
13:22:40 ameeda Thanks !
13:23:01 jaypipes np
13:35:36 zioproto sdague: sorry yesterday I had to leave and I did not answer. The version of Openstack is Newton
13:39:47 ameeda jaypipes: sorry for inconvenience, should I build new server from snapshot image ?, I can't find any details can help me when I use api.get_server(created_server['id']) , please check the log to see what I got http://paste.openstack.org/show/646510/
13:46:48 ameeda jaypipes: ?
13:48:46 edmondsw gibi we respun https://review.openstack.org/#/c/422512/ for a comment from stephenfin and lost your +2. Please take another look when you get a chance. Thanks!
13:50:37 edmondsw stephenfin do you think you'll be able to look at the SEA patch? It's similar to OVS, builds on that to add another type of networking: https://review.openstack.org/#/c/523216/
13:53:54 gibi edmondsw: it is on my list for today
13:54:05 edmondsw gibi great, tx
13:59:57 bauzas gibi: could I beg you for reviewing https://review.openstack.org/#/c/528832/10 ? ;)
14:00:26 bauzas FWIW, there will be a GPU track for the OpenStack Summit, and I'd love to showcase with Xen folks ;)
14:07:51 gibi bauzas: I will try to look at it, but no promises for today
14:08:04 bauzas np, I totally understand :)
14:08:11 bauzas I basically have the same answer
14:09:19 gibi bauzas: btw, multiattach also need a second core https://review.openstack.org/#/c/267587 :)
14:09:36 ildikov gibi: I typed it in, but then I thought I will not put you in a mood where you just throw it to the end of your list :)
14:09:49 ildikov gibi: so I haven't sent the message
14:09:53 gibi ildikov: no worries :)
14:10:28 ildikov gibi: yeah, I'm trying my best not to worry about it... :)
14:10:58 bauzas gibi: just sayin'
14:11:04 bauzas (15:07:51) gibi: bauzas: I will try to look at it, but no promises for today
14:11:10 bauzas same for multiattach :p
14:11:15 gibi bauzas: understood :p
14:11:25 ildikov gibi: I thought to ask jaypipes or melwitt to look at the chain as it's not terribly long and Matt touched all of them so it can't be terribly wrong either :)
14:11:27 bauzas why for the hell all the things I need to do happen at the same time ?
14:12:02 gibi bauzas: feels like a deadline comming up
14:12:03 ildikov bauzas: I believe it's called code freeze? :)
14:12:49 stephenfin sean-k-mooney: That os-vif patch is +Wd now. Thanks for the review, jaypipes :)
14:13:26 stephenfin bauzas: I'm assuming you were linking the vGPU patch earlier? Won't get to it today, but I'll star it for review later this week :)
14:13:39 stephenfin edmondsw: Ditto for your SEA patch ^
14:13:58 bauzas stephenfin: yup, see gibi's harassement ping I made to him
14:14:02 edmondsw stephenfin tx
14:14:24 bauzas ildikov: gibi: the multiattach patch doesn't seem crazy to review
14:14:40 ildikov bauzas: it's a chain
14:14:40 bauzas I'll try to look at it today if I have time
14:14:52 bauzas I know
14:14:56 ildikov bauzas: there's one longer patch but it's mostly tests
14:15:10 ildikov bauzas: would be a huge help, thanks in advance!
14:15:17 bauzas I first need to do internal stuff, but 'll try
14:15:49 ildikov bauzas: sure, I know how that goes :)
14:16:28 bauzas the fact that internal stuff becomes prioritary litterally one week before FF makes me puzzled, but yeah, that's life
14:16:29 gibi bauzas: I still have to look at the api part of the multiattach but what I saw is not worrying
14:19:07 openstackgerrit Claudiu Belu proposed openstack/nova master: tests: fixes mock autospec usage https://review.openstack.org/447505
14:23:14 gibi edmondsw: I +W-d the powerVM os-vif. Nice work!
14:26:59 edmondsw gibi thanks!
14:28:14 jaypipes ameeda: sorry, went breakfasting
14:29:03 jaypipes ameeda: yeah, I see you getting 'metadata': {}
14:29:14 jaypipes ameeda: try running as admin instead of a normal user.
14:29:15 cdent at long last the monster rises from the crypt in search of food
14:29:28 jaypipes cdent: rooooaaarrr!
14:30:02 cdent jaypipes: you see my bloggy thing about placement container experiment from yesterday?
14:30:17 jaypipes bauzas: "why for the hell all the things I need to do happen at the same time ?" <-- I don't... I don't even know what that means :)

Earlier   Later