13:59:22 #startmeeting glance 13:59:22 Meeting started Thu Apr 13 13:59:22 2023 UTC and is due to finish in 60 minutes. The chair is pdeore. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:59:22 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:59:22 The meeting name has been set to 'glance' 13:59:22 #topic roll call 13:59:22 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 13:59:25 o/ 14:00:05 o/ 14:00:37 I think cyril is not there 14:01:00 let's wait few more mins for othersto join 14:01:15 o/ 14:01:37 ok lets' start 14:01:39 #topic Updates 14:01:57 Abhishek has submitted Bobcat PTG summary, thanks abhishekk! 14:01:59 #link https://lists.openstack.org/pipermail/openstack-discuss/2023-April/033187.html 14:02:16 np! 14:02:27 Regarding upcoming In Person PTG , I missed to submit the survey which is to book slots for PTG 14:02:42 So before contacting to ptg@openinfra.dev I want to know will there be anyone to represent glance team if required? 14:03:04 abhishekk was ready to travel but due to visa issue no one from India would be able to travel 14:03:28 rosmaita, are you going to travel for this ptg? 14:03:55 possibly 14:04:01 if noone from a project is going to be there, I don't see any point in contacting the Kendall for the same 14:04:06 ohh gr8 14:04:19 rosmaita, Is it possible for you to represent the glance team if required? 14:04:31 sure, i can do that 14:04:47 Great, Thanks ! 14:05:29 moving to next .. 14:05:35 #topic release/periodic jobs updates 14:05:56 M1 is 4 weeks away from now 14:06:28 Periodic job all green today except some failure on functional tests, but i think that's also because of the PaseDeploy issue 14:06:41 which is fixed now 14:07:19 next topic i had added regarding the same CI issue but since it's fixed now, skipping this topic 14:07:39 moving to next 14:07:51 #topic Any Important Bugs for this week 14:08:12 As discussed during PTG, I have added this topic to discuss the imp bugs we have which cyril was going to update 14:08:33 but since he is not around today I think let's discuss that in next week's meeting 14:08:46 (sorry, i am in another meeting, please ping me when you get to open discussion) 14:08:59 rosmaita, sure 14:09:33 rosmaita, we are moving to open Discussions 14:09:36 #topic Open Discussions 14:09:46 :) 14:09:50 that was quick! 14:10:22 to follow up about the summit/PTG, is no one from glance (other than me) going to be there? 14:10:26 yeah had to skip 2 topics due to no specific updates :) 14:10:45 probably no one 14:10:46 well, only abhishekk was ready to travel but due to visa issue no one from India would be able to travel 14:10:55 :( 14:11:10 I can assist you virtually if required 14:12:04 ok, let's talk later about what we want to do 14:12:20 pdeore: i guess you were asking do you need to reserve space for glance? 14:13:16 pdeore, has electricity failure 14:13:26 she will be back in couple of minutes 14:13:37 ok, i can wait, no rush 14:13:57 yeah, I think we don't need to reserve the space 14:14:11 btw, i think dansmith will be at the summit, i believe it is real close to where he lives 14:14:19 I will not 14:14:25 :( 14:14:40 oh sorry, I missed that this is now in my timezone again, sorry! 14:14:48 I never met dansmith in person :( 14:15:03 abhishekk: consider yourself luck! 14:15:07 *lucky 14:15:10 :D 14:15:23 * abhishekk shuts up 14:15:43 my first meeting with dansmith i am sure he doesn't remember, was at a nova session at an old-style summit, he slapped me down about a comment i made about microversions 14:15:47 it is fairly closeish to where I live, but there is an international border between us :) 14:16:11 rosmaita: I remember that conversation, but I surely don't remember a smackdown 14:16:35 i was young and fragile in those days 14:16:39 I missed that as well :D 14:17:08 rosmaita, any suggestions on multihash thing? 14:17:14 rosmaita: well, sorry if I was young and poorly behaved back then :) 14:17:35 abhishekk: what's the multihash thing? 14:17:53 multiple checksum on open discussion 14:17:54 sorry I lost the electricity due to thundering & lightening here 14:18:01 https://etherpad.opendev.org/p/glance-team-meeting-agenda 14:18:45 no worries, we were just revisiting old memories 14:19:32 so create the image, set the expected hash, and let glance check it after it has webdownloaded it? 14:19:33 ahh cool then ..:) 14:20:01 kind of something 14:21:00 i guess that's fine, as long as people don't use a weird algorithm 14:21:09 i'm thinking of download 14:21:11 at the moment I think we have hashing_algorithm config which is common for all the images, is he talking about setting hash algo for each image so that we can provide different algo? 14:21:34 iirc the client reads it out of the os_hash_algo field, and instantiates the validator 14:21:37 presumably it would have to be a confined list of supported algorithms 14:22:05 yeah, would have to be recognized by (i forget what the python module name is, we have it documented) 14:22:09 abhishekk: I'm thinking you have to say "I want you to check the SHA384 of this image when you download it" 14:22:22 yeah 14:22:34 you could still let glance hash it however it expects for the actual storage, but use another hash value and algo to validate before it agrees to eat it 14:23:45 but that should be part of the client (which you are using for download), right? 14:24:13 no, they want it for web download right? 14:24:22 so the case where the user never gets to check the image themselves 14:24:30 so they want to make sure glance downloaded the right thing 14:24:35 yeah, so atm how you can ask it to verity with different algo 14:24:49 unless we add additional task for it 14:25:03 if they know the SHA384, and glance will only calculate the SHA256 for them, they can't tell that they're booting the thing they instructed 14:25:17 abhishekk: yeah you'd need to add functionality for this, isn't that the proposal? some new feature? 14:25:25 correct 14:25:25 well, they could add the "extra" hash as an image property 14:25:29 we are on same page now 14:25:35 it just wouldn't be automated 14:25:35 rosmaita: right exactly 14:25:54 expected_hash=SHA384:2139134undfidsf;ih 14:25:55 but they want to verify it as well 14:26:35 well, they say upload too, yeah, but that would work the same right? 14:26:50 "here's the sha384 of the thing I'm about to upload, check it when I'm done" 14:27:04 Probably if we agree that this is a good addition then let them ask to provide spec 14:27:26 I think the use case is probably more useful for web-download, 14:27:52 but I guess if you don't know what hash glance will use on upload, you might want to have this to avoid needing to upload, see the hash, and then re-check locally 14:28:14 agree 14:29:02 yeah, for web download, it seems like you want glance to check something to make sure it got the right thing, but you don't need glance to actually store that hash (unless you don't trust glance, i guess) 14:29:24 right, but unless glance deletes that key, 14:29:38 if it agrees to store it then the hash is recorded by virtue of it having matched 14:31:33 i guess the current config could be used as the default os_hash_algo, but there's no reason why we couldn't allow a user to specify an alternative 14:31:45 well, 14:31:57 you wouldn't have uniformity, but i guess you don't really need it 14:32:02 I don't think we need to change the os_hash_algo right/ 14:32:35 you'd have os_hash_algo=SHA1 os_hash_value=$sha1hash and also requested_hash=SHA384:$sha384hash 14:32:41 ++, we just need to have additional api parameter as input or extra_property as suggested by dansmith 14:33:08 and if the web-download completed and you went to active, then you know the requested_hash must have matched and you could use it or the sha1 default if you want to validate right? 14:33:36 correct 14:34:01 (i am in 2 meetings and doing a bad job of following the discussion in both) 14:34:17 well, we can discuss on a spec anyway :) 14:35:02 right, I think lets suggest him that we are ok with Idea, and need a spec for further strengthening the same 14:35:15 agree, it does seem useful, and actually we should be doing something like this for web-download anyway 14:35:26 yeah 14:35:44 cool 14:36:00 pdeore, if you are back, I guess we can move forward 14:36:34 * abhishekk we can not end meeting unless pdeore is back :D 14:36:45 yeah, so we are going to suggest him to propose a spec for this , right? 14:36:59 I have lost so many messages 14:37:17 yeah 14:37:27 ok, I will go thro the logs later 14:37:30 yes 14:37:37 I have jst added one line in agenda 14:37:42 so yeah that's it for today I guess 14:37:47 you can later add the link of logs 14:37:55 abhishekk, ohh thanks ! 14:38:01 just to remind you 14:38:08 one other thing from me 14:38:08 my question is if we're not going to have a glance table at the PTG, what do you need me to do to represent glance? 14:38:37 you need to followup with Artem 14:38:52 rosmaita, if there is any specific session for project updates 14:39:43 or if they want to record some videos (like in the past) for project updates 14:39:44 abhishekk, yeah I have already informed him but will again ping him 14:39:53 cool, thanks 14:40:04 that's it from me 14:40:04 i think they did the project updates already at openstack live 14:40:58 then I don't think we need to represent anything else 14:41:22 so with Artem, is it about openstackclient, SDK? 14:42:23 yes 14:43:29 ok, i can coordinate offline with the Glance team before the PTG to get a sense of what our thoughts are 14:43:44 sounds good 14:44:14 i am really bummed that no one else from glance will be there, though 14:44:44 usually jokke_ travels but I am not sure this time he will be there as well 14:45:16 yeah, i have not talked to him, i'm just assuming he's not going 14:46:29 yep me too 14:47:02 Can we continue this discussion on glance channel 14:47:23 yep, I don't have anything else to add 14:47:29 thank you all for joining 14:47:36 i'm facing network issues :/ so we can end the meeting here and continue on glance channel 14:47:51 ++ 14:48:00 Thanks everyone for joining ! and sorry for the inconvenience 14:48:13 #endmeeting 14:48:37 you need to type endmeeting again 14:49:02 #endmeeting