| *** ducttape_ has joined #openstack-glance | 00:14 | |
| *** ducttape_ has quit IRC | 00:15 | |
| *** ctrath has joined #openstack-glance | 00:15 | |
| *** ducttape_ has joined #openstack-glance | 00:15 | |
| *** btully has quit IRC | 00:19 | |
| *** ctrath has quit IRC | 00:20 | |
| *** btully has joined #openstack-glance | 00:21 | |
| *** ducttape_ has quit IRC | 00:40 | |
| *** mtanin___ has joined #openstack-glance | 00:52 | |
| *** mtanino has quit IRC | 00:53 | |
| *** markvoelker has joined #openstack-glance | 00:55 | |
| *** btully has quit IRC | 00:56 | |
| *** btully has joined #openstack-glance | 00:57 | |
| *** liwei has joined #openstack-glance | 00:58 | |
| *** bpokorny_ has joined #openstack-glance | 01:00 | |
| *** jamielennox is now known as jamielennox|away | 01:03 | |
| *** bpokorny has quit IRC | 01:04 | |
| *** bpokorny_ has quit IRC | 01:05 | |
| *** namnh has joined #openstack-glance | 01:07 | |
| *** ducttape_ has joined #openstack-glance | 01:09 | |
| *** lujinluo has quit IRC | 01:09 | |
| *** lujinluo has joined #openstack-glance | 01:10 | |
| *** ducttape_ has quit IRC | 01:13 | |
| *** ducttape_ has joined #openstack-glance | 01:16 | |
| *** lujinluo has quit IRC | 01:17 | |
| *** jamielennox|away is now known as jamielennox | 01:22 | |
| openstackgerrit | liwei proposed openstack/glance: Add more resource url in readme.rst https://review.openstack.org/365525 | 01:27 |
|---|---|---|
| openstackgerrit | liwei proposed openstack/glance: Correct url in doc source https://review.openstack.org/368168 | 01:31 |
| *** hemanthm|afk is now known as hemanthm | 01:39 | |
| *** btully has quit IRC | 01:40 | |
| *** btully has joined #openstack-glance | 01:41 | |
| *** ducttape_ has quit IRC | 02:01 | |
| openstackgerrit | Jin Li proposed openstack/glance-specs: change the example URLs in narrative-style API docs for glance-specs https://review.openstack.org/353787 | 02:05 |
| *** tsymanczyk has quit IRC | 02:13 | |
| *** mtanino has joined #openstack-glance | 02:16 | |
| openstackgerrit | Jin Li proposed openstack/glance: change the example URLs in api-ref for Glance https://review.openstack.org/353676 | 02:17 |
| *** mtanin___ has quit IRC | 02:17 | |
| *** ducttape_ has joined #openstack-glance | 02:19 | |
| *** mtanino has quit IRC | 02:22 | |
| *** TravT has joined #openstack-glance | 02:33 | |
| *** mugsie has quit IRC | 02:33 | |
| *** mugsie has joined #openstack-glance | 02:35 | |
| *** jamielennox is now known as jamielennox|away | 02:37 | |
| *** jamielennox|away is now known as jamielennox | 02:51 | |
| openstackgerrit | Tony Xu proposed openstack/glance: Fixed a W503 pep8 issue https://review.openstack.org/375204 | 02:53 |
| *** ducttape_ has quit IRC | 03:01 | |
| openstackgerrit | liwei proposed openstack/glance_store: Remove debtcollector in requirements.txt https://review.openstack.org/375206 | 03:03 |
| *** btully has quit IRC | 03:04 | |
| *** david-lyle has quit IRC | 03:04 | |
| *** btully has joined #openstack-glance | 03:05 | |
| *** btully has quit IRC | 03:11 | |
| *** btully has joined #openstack-glance | 03:12 | |
| *** prateek has joined #openstack-glance | 03:19 | |
| *** mtanino has joined #openstack-glance | 03:26 | |
| *** namnh has quit IRC | 03:31 | |
| *** mtanino has quit IRC | 03:41 | |
| *** sdake has quit IRC | 04:02 | |
| *** GB21 has joined #openstack-glance | 04:10 | |
| *** prateek has quit IRC | 04:47 | |
| *** flwang1 has quit IRC | 04:55 | |
| *** bhagyashri has quit IRC | 05:01 | |
| *** sabari has quit IRC | 05:17 | |
| *** pdeore has joined #openstack-glance | 05:28 | |
| *** bhagyashri has joined #openstack-glance | 05:35 | |
| *** sabari has joined #openstack-glance | 05:36 | |
| *** lujinluo has joined #openstack-glance | 05:40 | |
| *** groen692 has joined #openstack-glance | 05:51 | |
| *** groen692 has quit IRC | 05:52 | |
| *** groen692 has joined #openstack-glance | 05:55 | |
| *** prateek has joined #openstack-glance | 05:57 | |
| *** prateek has quit IRC | 06:01 | |
| *** prateek has joined #openstack-glance | 06:01 | |
| *** rcernin has joined #openstack-glance | 06:15 | |
| *** aswadr_ has joined #openstack-glance | 06:25 | |
| *** GB21 has quit IRC | 06:45 | |
| *** rrecio_ has quit IRC | 06:45 | |
| *** GB21 has joined #openstack-glance | 06:57 | |
| *** ejat has joined #openstack-glance | 07:05 | |
| ejat | hi .. what is the replacement for --copy-from option in image v2 ? | 07:06 |
| openstackgerrit | liwei proposed openstack/glance: Rename get_repo to get_image_repo in gateway.py https://review.openstack.org/362679 | 07:26 |
| openstackgerrit | liwei proposed openstack/glance: Change SafeConfigParser into ConfigParser https://review.openstack.org/368405 | 07:27 |
| openstackgerrit | liwei proposed openstack/glance: Add programming language python3 in classifier https://review.openstack.org/353905 | 07:28 |
| *** dshakhray has joined #openstack-glance | 07:29 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-glance | 08:00 | |
| *** dshakhray has quit IRC | 08:08 | |
| Jokke_ | ejat: there isn't, at least yet | 08:30 |
| openstackgerrit | YAMADA Hideki proposed openstack/glance_store: Sheepdog: fix command execution failure https://review.openstack.org/365564 | 08:39 |
| *** GB21 has quit IRC | 08:42 | |
| *** GB21 has joined #openstack-glance | 08:46 | |
| *** lujinluo has quit IRC | 08:57 | |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** openstackgerrit has joined #openstack-glance | 09:03 | |
| *** GB21 has quit IRC | 09:06 | |
| *** dshakhray has joined #openstack-glance | 09:20 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/python-glanceclient: Send HTTP headers with hyphens to API server https://review.openstack.org/375357 | 09:42 |
| *** jaosorior has joined #openstack-glance | 09:47 | |
| *** GB21 has joined #openstack-glance | 09:49 | |
| *** Dinesh_Bhor has joined #openstack-glance | 09:54 | |
| *** TravT has quit IRC | 10:04 | |
| *** liwei has quit IRC | 10:15 | |
| *** GB21 has quit IRC | 10:53 | |
| jaosorior | Jokke_: hey, so I ended up stumbling on a bug that you were assigned to https://bugs.launchpad.net/glance/+bug/1276887 and took a stab at it https://review.openstack.org/#/c/375357/ can you check it out? | 11:07 |
| openstack | Launchpad bug 1276887 in Glance "Glance uses underscores in header names - stripped by many proxies" [High,Opinion] - Assigned to Erno Kuvaja (jokke) | 11:07 |
| *** ccesario has joined #openstack-glance | 11:09 | |
| *** smatzek has joined #openstack-glance | 11:12 | |
| Jokke_ | jaosorior: will do | 11:17 |
| *** smatzek has quit IRC | 11:17 | |
| jaosorior | Jokke_: thanks | 11:17 |
| *** aswadr_ has quit IRC | 11:22 | |
| *** sabari has quit IRC | 11:56 | |
| *** zul has quit IRC | 12:07 | |
| *** GB21 has joined #openstack-glance | 12:07 | |
| *** pdeore has quit IRC | 12:08 | |
| *** catintheroof has quit IRC | 12:08 | |
| *** sabari has joined #openstack-glance | 12:23 | |
| *** prateek has quit IRC | 12:26 | |
| sigmavirus | jamielennox: so there's no good replacement then? | 12:28 |
| *** zul has joined #openstack-glance | 12:32 | |
| *** rcernin has quit IRC | 12:38 | |
| *** takedakn has quit IRC | 12:42 | |
| *** ducttape_ has joined #openstack-glance | 12:44 | |
| *** pcaruana|afk| has joined #openstack-glance | 12:44 | |
| *** ducttape_ has quit IRC | 12:45 | |
| *** ducttape_ has joined #openstack-glance | 12:46 | |
| *** sdake has joined #openstack-glance | 12:48 | |
| *** GB21 has quit IRC | 12:54 | |
| *** david-lyle has joined #openstack-glance | 12:56 | |
| *** prateek has joined #openstack-glance | 12:58 | |
| *** pcaruana|afk| has quit IRC | 13:00 | |
| *** pcaruana has quit IRC | 13:01 | |
| *** pcaruana has joined #openstack-glance | 13:02 | |
| *** rcernin has joined #openstack-glance | 13:05 | |
| *** prateek_ has joined #openstack-glance | 13:06 | |
| *** prateek has quit IRC | 13:07 | |
| *** e0ne has joined #openstack-glance | 13:11 | |
| *** jaosorior has quit IRC | 13:13 | |
| *** e0ne has quit IRC | 13:14 | |
| *** jaosorior has joined #openstack-glance | 13:14 | |
| *** mtanino has joined #openstack-glance | 13:19 | |
| *** parora has joined #openstack-glance | 13:22 | |
| *** e0ne has joined #openstack-glance | 13:22 | |
| *** prateek_ has quit IRC | 13:24 | |
| *** ctrath has joined #openstack-glance | 13:25 | |
| *** prateek_ has joined #openstack-glance | 13:25 | |
| *** rcernin has quit IRC | 13:26 | |
| *** parora has quit IRC | 13:27 | |
| *** parora has joined #openstack-glance | 13:28 | |
| *** prateek_ has quit IRC | 13:32 | |
| *** pdeore has joined #openstack-glance | 13:34 | |
| *** ducttape_ has quit IRC | 13:36 | |
| *** ekarlso_ has joined #openstack-glance | 13:37 | |
| *** kragniz has quit IRC | 13:48 | |
| *** kragniz has joined #openstack-glance | 13:48 | |
| *** rcernin has joined #openstack-glance | 13:52 | |
| *** e0ne has quit IRC | 13:53 | |
| *** ctrath has quit IRC | 13:55 | |
| *** hussain_c has quit IRC | 13:56 | |
| *** ctrath has joined #openstack-glance | 13:58 | |
| *** ducttape_ has joined #openstack-glance | 14:02 | |
| *** ducttape_ has quit IRC | 14:06 | |
| *** parora has quit IRC | 14:06 | |
| *** ctrath has quit IRC | 14:19 | |
| *** ctrath has joined #openstack-glance | 14:24 | |
| *** TravT has joined #openstack-glance | 14:25 | |
| *** pdeore has quit IRC | 14:28 | |
| openstackgerrit | Hemanth Makkapati proposed openstack/glance: cpu and address space limitations on qemu-img https://review.openstack.org/375526 | 14:36 |
| *** pdeore has joined #openstack-glance | 14:48 | |
| *** rrecio has joined #openstack-glance | 14:48 | |
| *** rrecio_ has joined #openstack-glance | 14:50 | |
| *** pdeore has quit IRC | 14:52 | |
| *** pdeore has joined #openstack-glance | 14:53 | |
| *** rrecio has quit IRC | 14:53 | |
| *** burgerk has joined #openstack-glance | 14:59 | |
| *** pdeore has quit IRC | 15:00 | |
| *** jistr is now known as jistr|call | 15:03 | |
| *** ducttape_ has joined #openstack-glance | 15:07 | |
| *** rvasilets___ has joined #openstack-glance | 15:07 | |
| *** ducttape_ has quit IRC | 15:12 | |
| *** burgerk has quit IRC | 15:13 | |
| *** rcernin has quit IRC | 15:19 | |
| *** Dinesh_Bhor has quit IRC | 15:24 | |
| *** jistr|call is now known as jistr | 15:26 | |
| *** jistr is now known as jistr|biab | 15:28 | |
| *** bhagyashri has quit IRC | 15:30 | |
| sigmavirus | rosmaita: meeting about release things? | 15:31 |
| rosmaita | hello release team! | 15:31 |
| *** groen692 has quit IRC | 15:31 | |
| rosmaita | sigmavirus: too quick for me! | 15:31 |
| rosmaita | Jokke_: yt? | 15:31 |
| sigmavirus | flaper87: and nikhil too, right? | 15:32 |
| rosmaita | nikhil: hello! | 15:32 |
| rosmaita | not sure flaper87 can make it, he was traveling | 15:32 |
| rosmaita | i will update him via email | 15:32 |
| Jokke_ | rosmaita: yeah | 15:32 |
| nikhil | \o | 15:33 |
| rosmaita | Jokke_: hello | 15:33 |
| rosmaita | nikhil: o/ | 15:33 |
| nikhil | :) | 15:33 |
| Jokke_ | I have not seen anything coming from the proposal bot | 15:33 |
| rosmaita | ok | 15:33 |
| rosmaita | flaper87: yt? (just in case) | 15:34 |
| sigmavirus | I suspect he'd have answered me | 15:34 |
| Jokke_ | so I'm assuming we're good. OFC I don't think we're in crazy rush so I can double check Mon morning again | 15:34 |
| rosmaita | cool | 15:34 |
| rosmaita | so, this is for my edification | 15:34 |
| * sigmavirus has forgotten what we're looking for | 15:34 | |
| rosmaita | i haven't paid as close attention as i should, until it's necessarey | 15:35 |
| Jokke_ | sigmavirus: translations warranting RC2 | 15:35 |
| rosmaita | right, that's what we're looking for | 15:35 |
| Jokke_ | ar delaying it | 15:35 |
| rosmaita | other possibility would be late-breaking security bug into ocata, would ahve to be backported to stable/newton? | 15:36 |
| *** jistr|biab is now known as jistr | 15:36 | |
| Jokke_ | well that. Or something other seriously critical bug wise | 15:36 |
| rosmaita | ok | 15:36 |
| rosmaita | and then there's a possible RC3 also, is that correct? | 15:37 |
| sigmavirus | right | 15:37 |
| Jokke_ | as long as time permits we "can" roll as many as we need | 15:37 |
| rosmaita | ok | 15:37 |
| sigmavirus | Yeah although the Release managers might not appreciate it unless they're really critical | 15:37 |
| Jokke_ | obviously something is seriously wrong if we need to roll 4-5 RCs | 15:37 |
| rosmaita | gotcha | 15:37 |
| sigmavirus | So like there's a change to stable/newton proposed to stabilize how we sort drivers in our config file and I'm not sure that's suitable yet | 15:38 |
| * sigmavirus was hoping Jokke_ would weigh in ;) | 15:38 | |
| sigmavirus | Jokke_: https://review.openstack.org/365058 | 15:38 |
| sigmavirus | Oh, I also fast-tracked a translations patch a couple days ago fwiw | 15:39 |
| * sigmavirus was under the impression stable translation patches can get fast-tracked with one core | 15:39 | |
| Jokke_ | sigmavirus: yeah | 15:39 |
| rosmaita | sigmavirus: that sounds fine | 15:39 |
| * sigmavirus had only seen one prior to this one | 15:39 | |
| Jokke_ | sigmavirus: was that backported and merged to Newton? | 15:39 |
| Jokke_ | what comes to https://review.openstack.org/#/c/365058/ I'd tempted to -1 it | 15:40 |
| rosmaita | sigmavirus: do you have a link to the translastions patch? (just curious to see what it looks like) | 15:40 |
| sigmavirus | Jokke_: the translations patch? it was sent by the proposal bot | 15:40 |
| sigmavirus | rosmaita: hold please | 15:40 |
| Jokke_ | I don't actually mind getting it in, but I'd like to see up to date config files with it | 15:40 |
| Jokke_ | sigmavirus: yeah so it was targeted to master or to stable/newton? | 15:40 |
| sigmavirus | https://review.openstack.org/374041 | 15:41 |
| sigmavirus | stable/newton | 15:41 |
| Jokke_ | cool so that's fine then | 15:41 |
| Jokke_ | no need to worry about it | 15:41 |
| sigmavirus | seems to be the only one with that change-id | 15:41 |
| nikhil | https://review.openstack.org/#/q/project:openstack/glance+branch:stable/newton | 15:42 |
| *** parora has joined #openstack-glance | 15:42 | |
| nikhil | so, we definitely need a rc2 | 15:43 |
| Jokke_ | yeah | 15:43 |
| Jokke_ | In principle I'm against releasing anything at Fri but that isn't my call | 15:43 |
| Jokke_ | ;) | 15:43 |
| nikhil | as per http://lists.openstack.org/pipermail/openstack-dev/2016-September/104373.html | 15:43 |
| nikhil | we need the rc2 before sept 29 | 15:43 |
| nikhil | I think we'd delay it until that date and get as much as we can | 15:44 |
| nikhil | rather than having to spin rc3 | 15:44 |
| rosmaita | yeah, i would like to see https://review.openstack.org/371618 get in | 15:44 |
| Jokke_ | how about I merge any translations that might come in by/at Monday and we request RC2 after merge/when sufficient people are awake/online? | 15:44 |
| nikhil | I saw rc2 from Ceilometer and trove and trove-dashboard already, but that's just a reference | 15:45 |
| rosmaita | dumb question ... how does this work? we merge, tag, and then request release team to cut rc2 ? | 15:45 |
| nikhil | the above doesn't need us to rush the rc2 (but aware that community is moving fwd) | 15:45 |
| nikhil | rosmaita: no, it's much straightforward than that | 15:46 |
| nikhil | Jokke_: I was thinking wednesday for rc2 though | 15:46 |
| nikhil | rosmaita: just need to specify the information as a gerrit review against release repo | 15:46 |
| *** edmondsw has quit IRC | 15:46 | |
| *** ctrath has quit IRC | 15:46 | |
| Jokke_ | nikhil: that is just fine as well | 15:47 |
| *** ctrath has joined #openstack-glance | 15:47 | |
| Jokke_ | nikhil: I was more worried that you wanted one now ;) | 15:47 |
| nikhil | rosmaita: https://github.com/openstack/releases , the README is great. and you can use https://github.com/openstack/releases/blob/master/deliverables/newton/glance.yaml and https://github.com/openstack/releases/blob/master/deliverables/mitaka/glance.yaml and https://github.com/openstack/releases/blob/master/deliverables/mitaka/nova.yaml | 15:47 |
| Jokke_ | rosmaita: we do not tag | 15:47 |
| nikhil | as good references | 15:47 |
| Jokke_ | ^^ | 15:48 |
| rosmaita | ok, thanks | 15:48 |
| nikhil | Jokke_: ah | 15:48 |
| nikhil | rosmaita: the "hash" field and the "version" number are important | 15:48 |
| Jokke_ | so the tooling around releases repo has improved hugely as well, so that's not much manual work from the release management anymore. it's all around specially formatted yamls in the repo | 15:49 |
| sigmavirus | yep | 15:49 |
| *** parora has quit IRC | 15:49 | |
| Jokke_ | and if there is merge commit from the last commit, need to use the hash from the merge rather than the commit itself | 15:49 |
| *** edmondsw has joined #openstack-glance | 15:49 | |
| nikhil | rosmaita: only these lines will do https://github.com/openstack/releases/blob/master/deliverables/newton/glance.yaml#L59-L62 and some co-ordination with the release team (make sure to run "tox -e validate" before/in parallel when hitting git review) | 15:50 |
| *** MattMan has quit IRC | 15:50 | |
| * nikhil shuts up | 15:50 | |
| *** MattMan has joined #openstack-glance | 15:50 | |
| nikhil | Jokke_ is right | 15:50 |
| rosmaita | ok, so wed is one day before the deadline | 15:50 |
| nikhil | make sure you do a "git show" on latest stable/newton and use that hash | 15:51 |
| sigmavirus | rosmaita: re 371618, master hasn't merged adn I just re-reviewed that for you, sorry and you're welcome | 15:51 |
| rosmaita | maybe aim for 16:00 wed? or is the gate going to be really slow because everyone else has the same plan? | 15:51 |
| sigmavirus | rosmaita: no clue | 15:51 |
| nikhil | rosmaita: I'm okay with either, whoever is doing it needs to make sure the 4 release muskateers are aware of our plans. | 15:52 |
| nikhil | musketeers* | 15:53 |
| Jokke_ | rosmaita: the release jobs are fine and quick. and the Rel management takes care of the timing anyways | 15:53 |
| rosmaita | Jokke_: how does your wednesday look? | 15:53 |
| rosmaita | or is this sigmavirus 's thing (i am getting confused on who's responsible for what) | 15:53 |
| Jokke_ | 14:00 to 15:00 booked | 15:54 |
| nikhil | rosmaita: glance-release team, we look out for each other :) | 15:54 |
| Jokke_ | rosmaita: ping me at your Wed morning and we'll walk through it so you know the process for Ocata | 15:55 |
| nikhil | rosmaita: just a maybe-non-useful suggestion , if you wanted to open the mtg agenda for next week and mention rc2 tentative plans, it may hint fly-by reviewers to bring imp reviews to notice https://etherpad.openstack.org/p/glance-team-meeting-agenda | 15:55 |
| rosmaita | Jokke_: sounds good | 15:55 |
| Jokke_ | unless sigmavirus insists | 15:55 |
| Jokke_ | ;) | 15:55 |
| rosmaita | nikhil: maybe the ML? since wed is before the meeting? | 15:56 |
| nikhil | rosmaita: yeah both are ideal, out developers are fragmented. some read the ML on time and some prefer mtgs :) | 15:57 |
| nikhil | our* | 15:57 |
| sigmavirus | heh, I don't insist but I can help where necessary | 15:57 |
| nikhil | rosmaita: but it's your call, you run the things now so no one gets to question you. muhahahah. jk :) | 15:58 |
| sigmavirus | nikhil: rosmaita might need to borrow your iron fist ;) | 15:58 |
| nikhil | rosmaita: it's hard to keep cross posting, so one most convenient place should help ;-) | 15:58 |
| nikhil | sigmavirus: heh :D | 15:58 |
| nikhil | well the thing is... | 15:59 |
| rosmaita | ok, to summarize: our plan is to have rc2 ready on wed | 15:59 |
| *** tsymanczyk has joined #openstack-glance | 15:59 | |
| rosmaita | and it looks like we need a rc2 because there have been some translations merged, and possibly more trans | 15:59 |
| nikhil | if we provide options people will get confortable of sourcing from either places and in case of rush someone misses a cross post we have big issue. so, better to make it a habit to stick to one particular process. whichever one it may be. | 16:00 |
| rosmaita | and maybe some bugs | 16:00 |
| rosmaita | nikhil: understood ... my problem with the thurs meeting is that it occurs after the rc2 deadline we've set for ourselves | 16:00 |
| sigmavirus | I'd like to get the qemu-img review hemanthm is working on merged for rc2 but it's critical enough that it could be merged after newton is tagged too | 16:00 |
| tsymanczyk | buenos dias personas de glance, y bonita viernes | 16:00 |
| rosmaita | sigmavirus: ++ | 16:01 |
| nikhil | rosmaita: correct, whatever gets in gets in. i18n are already there so rc2 is mandatory. if nothing more gets in, we will have it (rc2) so. | 16:01 |
| rosmaita | tsymanczyk: hang on a minute, please | 16:01 |
| tsymanczyk | yeah np. | 16:01 |
| rosmaita | ok, so it sounds like we are ok. i will send out a ML update on RC2 after community images sync | 16:01 |
| rosmaita | thanks nikhil sigmavirus Jokke_ (and flaper87 wherever you are) | 16:02 |
| nikhil | rosmaita: yeah, I'm just saying don't make it your problem otherwise problems may pile on you. people are expected to read the final meeting agenda 24 hrs before thurs meeting anyway (re: best practices) | 16:02 |
| sigmavirus | yw rosmaita | 16:02 |
| rosmaita | nikhil: i appreciate the advice! | 16:02 |
| nikhil | :) | 16:03 |
| rosmaita | ok, thanks release team! | 16:04 |
| rosmaita | tsymanczyk: thanks for waiting | 16:04 |
| rosmaita | you have the floor | 16:04 |
| Jokke_ | :) | 16:05 |
| tsymanczyk | sweet. so for everyone's reference, https://review.openstack.org/#/c/369110/ is the change we're talking about. | 16:05 |
| tsymanczyk | it is the first of two changesets for community images - and its sole purpose in life is to replace the is_public db images column with an enum to allow for the new values. | 16:05 |
| tsymanczyk | overall i'm happy with the state of it (in my bubble), except for ONE remaining thing | 16:06 |
| tsymanczyk | for a lot of unit tests, i needed to change their success conditions from testing the dict attribute of is_public to visilibility | 16:06 |
| tsymanczyk | and for the little fake tests that create internal dicts (fixtures) and just modify them a little, that doesn't bother me. | 16:07 |
| tsymanczyk | but there are a few tests that go through a full spin of the interface | 16:07 |
| tsymanczyk | and (imo) it isn't okay that those tests needed to be altered to pass | 16:07 |
| tsymanczyk | it has the identical consequence to an end user who has code relying on the v1 interface returning the is_public header | 16:07 |
| tsymanczyk | they would currently break | 16:07 |
| Jokke_ | tsymanczyk: so we cannot break v1 to support this | 16:08 |
| Jokke_ | if needed we need some bubblegum on the v1 code to translate that visibility back | 16:08 |
| tsymanczyk | now going forward, once the second patch enabling the new shared and community visibility values gets in it isn't clear to me that v1 would be able to utilize these new values at all. and especially with the new 'shared' value, i'm not sure how that's going to work with v1. but this patch at least seems like it shouldn't break v1 | 16:09 |
| tsymanczyk | jokke_ agreed. | 16:09 |
| rosmaita | i think the idea is that community images are is_public == 0 | 16:09 |
| tsymanczyk | jokke_ my current thought is to just decorate the headers as they get emitted from the v1 api with an the additional is_public value | 16:09 |
| tsymanczyk | rosmaita : certainly. but i'm wondering without the v1 ability to set anything but true or false, how are they going to be able to specify shared / community ? | 16:10 |
| rosmaita | tsymanczyk: v1 users can't do it | 16:10 |
| rosmaita | it will be weird in installations that use both v1 and v2 | 16:10 |
| rosmaita | but now that nova uses v2, shouldn't be much of an issue, i don't think | 16:11 |
| tsymanczyk | rosmita : okay. how are you picturing "not breaking" v1 image sharing then? | 16:11 |
| tsymanczyk | because i have no concept atm. | 16:11 |
| rosmaita | well, current v1 sharing has is_public == 0 | 16:11 |
| nikhil | tsymanczyk: that's exactly I was referring to for compat layers besides any admin_as_user magic. | 16:11 |
| openstackgerrit | Dharini Chandrasekar proposed openstack/glance: Fixing inconsistency in Glance store names. https://review.openstack.org/359402 | 16:11 |
| rosmaita | tsymanczyk: i should prob shut up and read through this patch to get a better grasp of your concerns | 16:12 |
| nikhil | tsymanczyk: not breaking v1 will need some translation of the visibility understanding in terms of is_public but that is likely going to be different for different use cases. | 16:12 |
| tsymanczyk | rosmaita : not at all. my concerns for this patch have been resolved. once i fix the headers it'll be 100% back compat. the current issue is for the next patch - since the v2 workflow is going to change to have an explicit "shared" visibility value, and v1 relies on the implicit behaviour of is_public=0 -- this ... needs some thought. | 16:13 |
| nikhil | tsymanczyk: for example, the is_public flag is hacked for admin_as_user but is not for simply admin. | 16:13 |
| *** bkopilov has quit IRC | 16:14 | |
| rosmaita | tsymanczyk: i think i understand the problem | 16:14 |
| nikhil | anyway, I am sorta in this state where I've seen the patch details just enough to realize this is the right direction but still pondering conceptually of what the code needs to look like. | 16:15 |
| tsymanczyk | nikhil : thanks for reminding me. the last issue i wanted to bring up is that i'm going to need explicit instruction on how to utilize your kind guide in the etherpad. regarding the difference between user-admin and glance-admin -- my only understanding is when "this" user has the "admin" role. how that applies to user-admin and glance-admin and the | 16:15 |
| tsymanczyk | difference between the two is lost to me. | 16:15 |
| rosmaita | v1 sharing uses membership only, v2 will use memberhip + visibility | 16:15 |
| nikhil | this is certainly induction effect of markwash! | 16:15 |
| tsymanczyk | rosmaita : exactly. | 16:15 |
| tsymanczyk | there may be some gross hacky code until v1 dies. | 16:15 |
| rosmaita | so if we dont' adjust v1, images wiht members will show up as shared in v1 but not in v2 | 16:15 |
| rosmaita | gotcha | 16:15 |
| tsymanczyk | certainly nobody wants that, but as discussed earlier we certainly can't break v1. | 16:16 |
| rosmaita | i think gross hacky code on a deprecated api is acceptable (and i said that in a logged channel!) | 16:17 |
| tsymanczyk | :) | 16:17 |
| nikhil | tsymanczyk: yep yep yep! the gross hacky code needs to live outside of the regular modules please!!!! we need strong documentation of the hacks / why, when, where/ and keep them packed in a separate packing-box, just like we've moved the house but don't want to open that box! | 16:17 |
| rosmaita | in the release notes we can say, if you intend to use community images, we strongly suggest that you have a v2-only glance deployment | 16:17 |
| tsymanczyk | certainly. i have no interest in leaving mystery work for whoever succeeds me. | 16:18 |
| *** jistr is now known as jistr|afk | 16:18 | |
| nikhil | well...if hierarchical images is next, guess who's that!? | 16:18 |
| nikhil | :D | 16:18 |
| tsymanczyk | in even the most perfect of worlds, i can't picture how v1 would ever actaully allow the use of the community value. right now my goal is to not break shared behaviour. | 16:18 |
| tsymanczyk | santa claus? | 16:19 |
| rosmaita | tsymanczyk: that's the correct goal | 16:19 |
| tsymanczyk | when will v1 officially be flushed? | 16:19 |
| nikhil | tsymanczyk: totally on translating the membership+visibility understanding separately for v1 (hint hint: the compat layer) | 16:19 |
| nikhil | tsymanczyk: To be *really* honest, never. but optimistically, 2-4 years. | 16:20 |
| tsymanczyk | actually one more question / clarification. so the plan is yes to ensure "is_public" gets emitted by the v1 api, yes? And beyond that, should I trim out the visibility value? or is having that one new value also there cool and/or desired? | 16:21 |
| *** jaosorior has quit IRC | 16:21 | |
| nikhil | (we say pike http://docs.openstack.org/releasenotes/glance/newton.html#deprecation-notes , but that's just being hopeful) | 16:21 |
| Jokke_ | no we should not send visibility in v1 | 16:21 |
| rosmaita | tsymanczyk: we dont' emit visibility now, do we? (in v1 response) | 16:21 |
| rosmaita | Jokke_: ++ | 16:21 |
| nikhil | == Jokke_ | 16:21 |
| tsymanczyk | in the current state, (to the best of my understanding) we emit is_public in v1 but not visibility. then v2 we emit visibility but not is_public. | 16:22 |
| rosmaita | don't want to encourage people to hack something together | 16:22 |
| tsymanczyk | okay cool. i'll take care of both of those things then. | 16:22 |
| Jokke_ | thanks tsymanczyk! | 16:22 |
| rosmaita | tsymanczyk: ok, that was my understanding ... so no visibility for v1 ! | 16:22 |
| tsymanczyk | thank you jokke_. | 16:22 |
| rosmaita | if you want visibility, you must upgrade to v2 | 16:22 |
| rosmaita | you == operators | 16:22 |
| Jokke_ | these two discussions back to back just made my Fri lots better ;) | 16:22 |
| tsymanczyk | so the only other blocker i have re this patch is instructions on user-admin / glance-admin, and how to do these in devstack. i really am lost on how to do a FULL manual test for these cases. | 16:23 |
| tsymanczyk | once those two things are cleared then i'll stop -1wf | 16:23 |
| nikhil | the admin that ships with devstack is the project/glance-admin | 16:24 |
| tsymanczyk | because of the 'admin' role? or being in the admin project? | 16:24 |
| nikhil | for user admin, we will need to set owner_is_tenant=False in glance-api.conf | 16:24 |
| nikhil | then see what the demo user's role is, I think it's not admin (we then need to create another user under the demo project namespace that has admin role for that tenancy). | 16:25 |
| nikhil | devstack has both project and user as demo, afaik | 16:25 |
| nikhil | 'being in the admin project' | 16:26 |
| nikhil | oops | 16:26 |
| nikhil | NO | 16:26 |
| nikhil | being admin "in" the project :) | 16:26 |
| rosmaita | yeah, i change the name of the project to 'demo_project' otherwise it gets too confusing | 16:27 |
| *** ctrath has quit IRC | 16:28 | |
| tsymanczyk | nikhil : so just to reiterate what you're saying. to test the project / glance-admin behaviour, it is only necessary to use the default 'admin' user within the admin project of a default devstack install. to test the user-admin behaviour, change owner_is_tenant=False in glance-api.conf, then assign the admin role to a user that is not 'admin'(username) ? | 16:28 |
| tsymanczyk | and not within the admin project? | 16:28 |
| nikhil | tsymanczyk: yes! | 16:28 |
| nikhil | and yes | 16:28 |
| tsymanczyk | perfect, thank you for being explicit. | 16:28 |
| tsymanczyk | i think that resolves everything i wanted to bring up this time then? hopefully by next glance meeting or commimg sync the first patch will be ready for reviews. | 16:29 |
| tsymanczyk | or perhaps already have some reviews on it. :) | 16:29 |
| nikhil | this is great news! | 16:29 |
| nikhil | but I'm packing and travelling in the coming days | 16:29 |
| tsymanczyk | anything else from anyone else? | 16:29 |
| tsymanczyk | nikhil : save travels! | 16:30 |
| nikhil | yeah, I may not be available for 2 weeks I think | 16:30 |
| nikhil | tsymanczyk: thanks! | 16:30 |
| rosmaita | tsymanczyk: sounds good, thanks for the update | 16:30 |
| tsymanczyk | anytime. thank you for the great feedback. | 16:30 |
| tsymanczyk | feedback / input | 16:30 |
| nikhil | I will try to think more about the compat layer design pattern as time permits. | 16:31 |
| nikhil | But I really think that you're DB change may not need it, we need to think about that for the next/community images API centric review. | 16:31 |
| Jokke_ | sigmavirus, rosmaita, hemanthm, nikhil: ref mailing list, sorry just did not want to let that go second time. | 16:31 |
| sigmavirus | Jokke_: hasn't come through to me yet | 16:32 |
| tsymanczyk | nikhil : agreed. the next patch will very likely need it. but my gut feel is that this one should be able to resolve all of the extremely valid points brought up without a compat layer. | 16:32 |
| nikhil | me neither | 16:33 |
| rosmaita | Jokke_: just catching up | 16:33 |
| nikhil | tsymanczyk: feel free to text me for a stuck patch on a +2 (will give you my number and personal email) | 16:33 |
| nikhil | tsymanczyk: would love to see this moving forward rather than a patch sitting with one +2 and waiting on the other onefor weeks. | 16:34 |
| tsymanczyk | nikhil : perfect, thank you. | 16:34 |
| sigmavirus | Jokke_: got it, I hope my response helps clarify my intent in the email too. This wasn't to call you out. | 16:41 |
| Jokke_ | sigmavirus: never taken as such, that one quote just jumped out second time and I hoped it was something I could just let slip ;) | 16:42 |
| *** dshakhray has quit IRC | 16:52 | |
| *** jistr|afk is now known as jistr | 16:59 | |
| *** markvoelker has quit IRC | 17:02 | |
| *** markvoelker has joined #openstack-glance | 17:02 | |
| openstackgerrit | Merged openstack/glance: Add note to docs on release notes prelude section https://review.openstack.org/365063 | 17:06 |
| *** ducttape_ has joined #openstack-glance | 17:08 | |
| *** catintheroof has joined #openstack-glance | 17:09 | |
| *** ducttape_ has quit IRC | 17:12 | |
| openstackgerrit | Alexander Bashmakov proposed openstack/glance-specs: Update ``updated_at`` field on image member change https://review.openstack.org/375629 | 17:14 |
| ejat | Jokke_: so how can i create image from URL into glance? | 17:19 |
| *** rvasilets___ has quit IRC | 17:28 | |
| *** ctrath has joined #openstack-glance | 17:32 | |
| sigmavirus | nikhil: rosmaita I think dharinic updated https://review.openstack.org/#/c/359402/27 asap for us so we could approve it. Given they adjusted the release notes for us and the rest looked good, care to +2+A? | 17:39 |
| openstackgerrit | Hemanth Makkapati proposed openstack/glance: CPU and address space limitations on qemu-img info https://review.openstack.org/375526 | 17:49 |
| *** ctrath has quit IRC | 17:51 | |
| *** ctrath has joined #openstack-glance | 17:55 | |
| *** btully has quit IRC | 17:55 | |
| *** hemanthm is now known as hemanthm|lunch | 17:55 | |
| *** btully has joined #openstack-glance | 17:56 | |
| dharinic | Thanks sigmavirus, hemanthm|lunch | 18:00 |
| sigmavirus | dharinic: good work on that btw | 18:01 |
| dharinic | Thanks :) | 18:01 |
| *** ducttape_ has joined #openstack-glance | 18:16 | |
| *** prateek has joined #openstack-glance | 18:23 | |
| *** e0ne has joined #openstack-glance | 18:29 | |
| *** ducttape_ has quit IRC | 18:31 | |
| *** e0ne has quit IRC | 18:43 | |
| *** e0ne has joined #openstack-glance | 18:47 | |
| *** hemanthm|lunch is now known as hemanthm | 18:50 | |
| *** dshakhray has joined #openstack-glance | 18:53 | |
| *** prateek has quit IRC | 18:55 | |
| openstackgerrit | Hemanth Makkapati proposed openstack/glance: CPU and address space limitations on qemu-img info https://review.openstack.org/375526 | 19:01 |
| rosmaita | hemanthm: you are too fast! | 19:01 |
| *** ckonstanski has joined #openstack-glance | 19:01 | |
| hemanthm | rosmaita: oops, were you reviewing it? | 19:02 |
| rosmaita | np | 19:02 |
| hemanthm | Jokke_: it's all good :) | 19:02 |
| rosmaita | hemanthm: left a question on your patch | 19:08 |
| hemanthm | looking | 19:09 |
| rosmaita | it may be out of scope | 19:09 |
| *** david-lyle_ has joined #openstack-glance | 19:10 | |
| hemanthm | rosmaita: let me look into what qemu-img convert is doing | 19:12 |
| rosmaita | thanks | 19:13 |
| *** david-lyle has quit IRC | 19:13 | |
| *** ctrath has quit IRC | 19:14 | |
| *** ctrath has joined #openstack-glance | 19:15 | |
| *** e0ne_ has joined #openstack-glance | 19:16 | |
| hemanthm | rosmaita: maybe #4. How about specifying the current image format with -f to the qemu-img convert command? | 19:17 |
| *** e0ne has quit IRC | 19:17 | |
| hemanthm | that will avoid the infering | 19:17 |
| *** e0ne_ has quit IRC | 19:17 | |
| *** catintheroof has quit IRC | 19:17 | |
| *** e0ne has joined #openstack-glance | 19:18 | |
| rosmaita | hemanthm: not sure ... is that info available? | 19:19 |
| *** e0ne has quit IRC | 19:19 | |
| *** e0ne_ has joined #openstack-glance | 19:19 | |
| rosmaita | also, i wonder whether it needs to read info out of the image even if it knows what the format is | 19:19 |
| *** e0ne_ has quit IRC | 19:19 | |
| hemanthm | rosmaita: in that case, we are doomed isn't it? | 19:19 |
| rosmaita | (i may have been misleading by focusing on the info call) | 19:19 |
| rosmaita | hemanthm: probably | 19:20 |
| rosmaita | i think we can do a probe by making the info call with limits | 19:20 |
| hemanthm | rosmaita: yeah, I'm open to doing that | 19:20 |
| *** e0ne has joined #openstack-glance | 19:20 | |
| rosmaita | but, we probably need to allow limits to be set for the actual conversion | 19:20 |
| rosmaita | in case something slips by with 1.9999999 CPU sec | 19:21 |
| *** e0ne has quit IRC | 19:21 | |
| rosmaita | but that seems a bigger change | 19:21 |
| hemanthm | yeah | 19:21 |
| rosmaita | maybe the probe + tasks are for admins is enough | 19:21 |
| hemanthm | alright, I'll look into adding the info call just before convert | 19:22 |
| rosmaita | thanks | 19:22 |
| hemanthm | or maybe we should reproduce the info attack with convert first | 19:22 |
| rosmaita | good idea | 19:22 |
| rosmaita | there are some recipes in the bug report | 19:23 |
| hemanthm | yeah | 19:23 |
| rosmaita | that's a good idea, my concern may be misplaced | 19:23 |
| hemanthm | this will probably apply to nova as well https://github.com/openstack/nova/blob/master/nova/virt/xenapi/vm_utils.py#L1128-L1129 | 19:26 |
| hemanthm | fungi ^ | 19:26 |
| *** burgerk has joined #openstack-glance | 19:27 | |
| *** ducttape_ has joined #openstack-glance | 19:32 | |
| *** ctrath has quit IRC | 19:32 | |
| fungi | hemanthm: can you summarize? i'm not quite sure i understand the suggestion | 19:33 |
| hemanthm | fungi: rosmaita pointed out that "qemu-img convert" may also try to infer the format of input image | 19:34 |
| *** ctrath has joined #openstack-glance | 19:34 | |
| hemanthm | so, is convert uses info internally, convert will also be susceptible to the same attack | 19:34 |
| hemanthm | to guard against it, we are planning to add an info call with process limits before we make the convert call | 19:35 |
| fungi | hemanthm: oh, that makes sense. can you mention that in the bug report? i'll see if we can get input from nova and cinder devs to see if they need to do the same | 19:36 |
| hemanthm | it's all a theory at this point, I'm trying to replicate the info attack with convert | 19:36 |
| fungi | good deal. i guess update the bug with that once you're more sure | 19:37 |
| hemanthm | sure | 19:37 |
| fungi | while i want to get the advisory out soon, i want even more not to have to issue a second one because we had an incomplete fix ;) | 19:37 |
| *** ducttape_ has quit IRC | 19:37 | |
| hemanthm | fungi: commented on the bug | 19:42 |
| *** thiagolib has joined #openstack-glance | 19:46 | |
| fungi | thanks! | 19:46 |
| *** burgerk has quit IRC | 19:53 | |
| *** sdake has quit IRC | 20:00 | |
| openstackgerrit | Merged openstack/glance: Fixing inconsistency in Glance store names. https://review.openstack.org/359402 | 20:05 |
| *** david-lyle_ is now known as david-lyle | 20:06 | |
| *** ducttape_ has joined #openstack-glance | 20:15 | |
| *** ducttape_ has quit IRC | 20:30 | |
| openstackgerrit | Nikhil Komawar proposed openstack/glance: Move deprecated store name mapping to resp. method https://review.openstack.org/375722 | 20:53 |
| *** ducttape_ has joined #openstack-glance | 20:54 | |
| *** ccesario has quit IRC | 21:06 | |
| *** ducttape_ has quit IRC | 21:07 | |
| *** kberger has quit IRC | 21:13 | |
| *** kberger has joined #openstack-glance | 21:13 | |
| *** thumpba has joined #openstack-glance | 21:31 | |
| *** ctrath has quit IRC | 21:44 | |
| *** thumpba has quit IRC | 21:46 | |
| *** kberger has quit IRC | 22:02 | |
| *** ducttape_ has joined #openstack-glance | 22:08 | |
| *** ducttape_ has quit IRC | 22:13 | |
| dharinic | Hey nikhil | 22:20 |
| *** rrecio_ has quit IRC | 22:20 | |
| nikhil | dharinic: \o | 22:20 |
| dharinic | I had a question regarding the bug for Swift multi tenancy https://bugs.launchpad.net/glance/+bug/1625075 | 22:20 |
| openstack | Launchpad bug 1625075 in Glance "Shared & public images no working with multi-tenant swift backend" [Critical,Triaged] - Assigned to Dharini Chandrasekar (dharini-chandrasekar) | 22:20 |
| dharinic | So as we were discussing the other day, a possible way of a fix would be to have the ACL set up.. | 22:20 |
| dharinic | I am still figuring a way to do this.. But I wanted to access the glance-db if we were to set the ACLs. I wasnt sure how to do that without editing requirements.txt | 22:22 |
| nikhil | yes, and I disagree with it :) | 22:22 |
| nikhil | I disagree with setting ACLs | 22:22 |
| dharinic | Hmmm.. kairat had mentioned we could fix it with ACLs. So I was just looking into that.. for which we might need access to glance-db | 22:23 |
| nikhil | but I also gave a reason why we shouldn't set ACLs! | 22:23 |
| dharinic | yess. I do remember that. :) | 22:24 |
| nikhil | and a pretty serious one, which has been ignored! | 22:24 |
| nikhil | (shame) | 22:24 |
| nikhil | not necessarily on you but on the entire set of people who managed to do so | 22:25 |
| dharinic | Ahhh not at all. I was just seeing how things work with ACLs (exploring) :D | 22:26 |
| nikhil | dharinic: http://eavesdrop.openstack.org/irclogs/%23openstack-glance/%23openstack-glance.2016-09-22.log.html#t2016-09-22T15:04:34 | 22:27 |
| dharinic | I do have that very link saved :) | 22:27 |
| dharinic | So the issue here is that the catalog from the request context is used for authentication.. https://github.com/openstack/glance_store/blob/master/glance_store/_drivers/swift/store.py#L1279-L1283 | 22:29 |
| dharinic | so even if the sharing is public/shared, the context.tenant is still scoped to the request | 22:30 |
| dharinic | I was just playing around with few tests locally with respect to this for which i thought I might access glance-db and see.. | 22:31 |
| dharinic | Hence the question | 22:31 |
| nikhil | dharinic: I don't have objections to accessing the glance-db | 22:32 |
| nikhil | dharinic: in fact, that's what I suggest | 22:33 |
| nikhil | dharinic: what we need is a pass by value of the sharing model to the multi tenant store logic | 22:33 |
| nikhil | dharinic: rather than setting the ACL everytime, you see a change in glance-db membership | 22:34 |
| nikhil | dharinic: the later being extremely racy, non-deterministic and an operation which does not provide consistency | 22:34 |
| nikhil | s/everytime, you see/everytime you see | 22:35 |
| dharinic | So we check the db for a change in membership and update the catalog every time there is a change? | 22:35 |
| nikhil | why do you need to update the catalog | 22:35 |
| nikhil | you cannot and should not | 22:35 |
| nikhil | dharinic: andrew's patch is just replacing the tenant value while authentication (not updating the catalog) | 22:36 |
| dharinic | Yeahh.. he fetches the storage_url and then replaces the tenant | 22:36 |
| dharinic | okayyy.. | 22:37 |
| dharinic | So coming back to the question of accessing the glance-db. How can I do it in glance_store? | 22:39 |
| dharinic | nikhil^ | 22:41 |
| nikhil | dharinic: you cannot and should not | 22:42 |
| nikhil | dharinic: we will need glance pass that info to the store | 22:42 |
| dharinic | I know I am missing something here.. | 22:43 |
| dharinic | Ahhhh | 22:43 |
| nikhil | you cannot (now) and should not (design in the future) | 22:43 |
| dharinic | gotcha | 22:43 |
| *** cdelatte has joined #openstack-glance | 22:43 | |
| dharinic | Thank you nikhil | 22:46 |
| nikhil | yw | 22:46 |
| nikhil | good work dharinic | 22:46 |
| nikhil | keep going like this! | 22:46 |
| dharinic | Thank you for sharing this bug with me. I am really enjoying working on this (though now I am just exploring things for this) | 22:47 |
| dharinic | Thank you :) | 22:47 |
| *** markvoelker has quit IRC | 22:47 | |
| *** thiagolib has quit IRC | 22:58 | |
| *** dshakhray has quit IRC | 23:01 | |
| *** tsymanczyk has quit IRC | 23:13 | |
| *** sdake has joined #openstack-glance | 23:17 | |
| *** ducttape_ has joined #openstack-glance | 23:18 | |
| *** ducttape_ has quit IRC | 23:22 | |
| *** ducttape_ has joined #openstack-glance | 23:41 | |
| *** ducttape_ has quit IRC | 23:45 | |
| nikhil | :) | 23:46 |
| *** markvoelker has joined #openstack-glance | 23:48 | |
| *** markvoelker has quit IRC | 23:52 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!