| *** haomaiwang has joined #openstack-swift | 00:01 | |
| *** chlong has joined #openstack-swift | 00:05 | |
| *** hogepodge has quit IRC | 00:16 | |
| *** hogepodge has joined #openstack-swift | 00:22 | |
| *** ianbrown has quit IRC | 00:23 | |
| *** ianbrown has joined #openstack-swift | 00:24 | |
| *** hogepodge has quit IRC | 00:27 | |
| *** dslevin_ has joined #openstack-swift | 00:35 | |
| notmyname | mattoliverau: I asked about the general christmas decorations that shops put up this time of year in Oz. you (IIRC) said that they used to do a lot more with the fake snow and stuff, but now it's more summer-ish | 00:42 |
|---|---|---|
| notmyname | mattoliverau: do you have all the holiday music piped into all the stores? especially the ones about cold and snow and stuff? | 00:42 |
| *** hogepodge has joined #openstack-swift | 00:43 | |
| notmyname | mattoliverau: or is there some australian holiday genre about "I'm dreaming of a sunny christmas" or "let it shine, let it shine, let it shine"? | 00:43 |
| mattoliverau | notmyname: some places still use fake snow, but mainly they just your xmas decoriations on the windows. | 00:44 |
| notmyname | but is it all about snowflakes and fat men in red overcoats? | 00:44 |
| mattoliverau | And yes, Christmas carols playing everywhere. There are some Australianised carols, but mostly just the normal ones. Inculding white christmas and frosty the snow man | 00:45 |
| notmyname | lol. ok | 00:45 |
| mattoliverau | lots of tinsel and boubles. | 00:46 |
| notmyname | evergreen fir trees | 00:46 |
| mattoliverau | notmyname: https://www.youtube.com/watch?v=r1aYkutjXfQ | 00:47 |
| notmyname | lol | 00:48 |
| notmyname | is that a real thing? I can't tell if that's a joke or not :-) | 00:49 |
| notmyname | I mean, I think it is, but... | 00:50 |
| notmyname | (a joke, that is) | 00:50 |
| notmyname | it also seems like something aussies would do (how's that for a stereotype! ;-) | 00:50 |
| mattoliverau | yeah, more of a joke, but its Australia (everything is a joke), so it'll still be played in stores on christmas (sometimes). | 00:51 |
| *** nadeem has joined #openstack-swift | 00:51 | |
| notmyname | lol | 00:52 |
| *** haomaiwang has quit IRC | 01:01 | |
| *** haomaiwang has joined #openstack-swift | 01:01 | |
| *** dslevin_ has quit IRC | 01:02 | |
| *** haomaiwang has quit IRC | 01:07 | |
| *** ianbrown has quit IRC | 01:11 | |
| *** ianbrown has joined #openstack-swift | 01:12 | |
| *** nadeem has quit IRC | 01:39 | |
| *** bkmz has quit IRC | 01:51 | |
| *** bkmz has joined #openstack-swift | 01:51 | |
| kota_ | hello | 01:52 |
| kota_ | Chrismas week? | 01:53 |
| notmyname | yeah. probably going to be a slower week this week and next | 01:53 |
| kota_ | notmyname: sounds good for enjoyining vacation :) | 01:54 |
| *** ianbrown has quit IRC | 01:57 | |
| *** ianbrown has joined #openstack-swift | 01:58 | |
| *** haomaiwang has joined #openstack-swift | 01:59 | |
| *** haomaiwang has quit IRC | 02:01 | |
| *** haomaiwa_ has joined #openstack-swift | 02:01 | |
| *** rminmin has joined #openstack-swift | 02:16 | |
| mattoliverau | kota_: o/ | 02:34 |
| *** ianbrown has quit IRC | 02:39 | |
| *** ianbrown has joined #openstack-swift | 02:43 | |
| *** sanchitmalhotra has joined #openstack-swift | 02:56 | |
| *** haomaiwa_ has quit IRC | 03:01 | |
| *** haomaiwang has joined #openstack-swift | 03:01 | |
| *** wanghua has quit IRC | 03:23 | |
| *** ianbrown has quit IRC | 03:26 | |
| *** links has joined #openstack-swift | 03:26 | |
| *** ianbrown has joined #openstack-swift | 03:26 | |
| *** asettle is now known as asettle-afk | 03:34 | |
| *** haomaiwang has quit IRC | 04:01 | |
| *** haomaiwang has joined #openstack-swift | 04:01 | |
| *** ianbrown has quit IRC | 04:10 | |
| *** ianbrown_ has joined #openstack-swift | 04:10 | |
| *** proteusguy_ has quit IRC | 04:23 | |
| *** mahatic has joined #openstack-swift | 04:26 | |
| *** haomaiwang has quit IRC | 04:33 | |
| *** proteusguy_ has joined #openstack-swift | 04:35 | |
| *** ppai has joined #openstack-swift | 04:46 | |
| *** trifon has joined #openstack-swift | 04:57 | |
| *** klrmn has quit IRC | 05:01 | |
| *** haomaiwa_ has joined #openstack-swift | 05:02 | |
| *** trifon has quit IRC | 05:03 | |
| *** nadeem has joined #openstack-swift | 05:11 | |
| *** ianbrown_ has quit IRC | 05:12 | |
| *** ianbrown_ has joined #openstack-swift | 05:12 | |
| *** nadeem has quit IRC | 05:17 | |
| *** nadeem has joined #openstack-swift | 05:18 | |
| *** SkyRocknRoll has joined #openstack-swift | 05:37 | |
| *** sanchitmalhotra has left #openstack-swift | 05:38 | |
| *** sanchitmalhotra has joined #openstack-swift | 05:39 | |
| *** nadeem has quit IRC | 05:44 | |
| *** SkyRocknRoll has quit IRC | 05:58 | |
| *** haomaiwa_ has quit IRC | 06:01 | |
| *** haomaiwang has joined #openstack-swift | 06:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:16 | |
| *** eranrom has joined #openstack-swift | 06:48 | |
| *** SkyRocknRoll has quit IRC | 06:50 | |
| *** ppai has quit IRC | 06:56 | |
| *** ChubYann has quit IRC | 06:59 | |
| *** haomaiwang has quit IRC | 07:01 | |
| *** haomaiwang has joined #openstack-swift | 07:01 | |
| *** SkyRocknRoll has joined #openstack-swift | 07:04 | |
| *** dslevin_ has joined #openstack-swift | 07:06 | |
| openstackgerrit | Peter Lisák proposed openstack/swift: Change schedule priority of daemon/server in config https://review.openstack.org/238799 | 07:06 |
| *** ppai has joined #openstack-swift | 07:08 | |
| *** wanghua has joined #openstack-swift | 07:08 | |
| *** dslevin_ has quit IRC | 07:11 | |
| *** chlong has quit IRC | 07:13 | |
| *** SkyRocknRoll has quit IRC | 07:23 | |
| openstackgerrit | Hisashi Osanai proposed openstack/swift: Fix duplication for headers in Access-Control-Expose-Headers https://review.openstack.org/258392 | 07:27 |
| *** SkyRocknRoll has joined #openstack-swift | 07:41 | |
| *** dslevin_ has joined #openstack-swift | 08:00 | |
| *** haomaiwang has quit IRC | 08:01 | |
| *** haomaiwang has joined #openstack-swift | 08:01 | |
| *** silor has joined #openstack-swift | 08:05 | |
| *** dslevin_ has quit IRC | 08:05 | |
| openstackgerrit | renminmin proposed openstack/python-swiftclient: Add default region name for swift SLI https://review.openstack.org/259879 | 08:06 |
| *** rledisez has joined #openstack-swift | 08:07 | |
| *** silor has quit IRC | 08:12 | |
| *** silor has joined #openstack-swift | 08:12 | |
| *** Jeffrey4l has quit IRC | 08:15 | |
| *** silor1 has joined #openstack-swift | 08:15 | |
| *** silor has quit IRC | 08:16 | |
| *** silor1 is now known as silor | 08:16 | |
| *** mac_ified has quit IRC | 08:17 | |
| *** Jeffrey4l has joined #openstack-swift | 08:22 | |
| openstackgerrit | renminmin proposed openstack/python-swiftclient: Add default region name for swift SLI https://review.openstack.org/259887 | 08:24 |
| *** chlong has joined #openstack-swift | 08:30 | |
| *** jsuchome has joined #openstack-swift | 08:33 | |
| *** geaaru has joined #openstack-swift | 08:40 | |
| openstackgerrit | Mahati Chamarthy proposed openstack/swift: Modify unit tests to include real crypto https://review.openstack.org/211451 | 08:50 |
| *** dslevin_ has joined #openstack-swift | 08:54 | |
| *** dslevin_ has quit IRC | 08:59 | |
| *** daemontool has joined #openstack-swift | 09:00 | |
| *** haomaiwang has quit IRC | 09:01 | |
| *** haomaiwang has joined #openstack-swift | 09:01 | |
| ntt | Hi, how can I check the used space in a swift cluster? | 09:13 |
| ntt | total used space, for all projects. | 09:13 |
| *** kei_yama_ has joined #openstack-swift | 09:16 | |
| *** kei_yama has quit IRC | 09:16 | |
| jsuchome | hello! could anyone review https://review.openstack.org/#/c/189258/ ? | 09:28 |
| *** openstack has joined #openstack-swift | 15:36 | |
| *** openstackstatus has joined #openstack-swift | 15:36 | |
| *** ChanServ sets mode: +v openstackstatus | 15:36 | |
| *** daddyjoseph97 has joined #openstack-swift | 15:44 | |
| *** SkyRocknRoll has joined #openstack-swift | 15:48 | |
| *** Jeffrey4l has quit IRC | 15:49 | |
| pdardeau | Good morning! | 15:53 |
| *** changbl has quit IRC | 15:54 | |
| *** daddyjoseph97 has quit IRC | 16:00 | |
| *** esker has quit IRC | 16:06 | |
| *** klrmn has joined #openstack-swift | 16:15 | |
| *** stevemar_znc is now known as stevemar | 16:28 | |
| *** gyee has joined #openstack-swift | 16:37 | |
| *** schizo is now known as bwall | 16:38 | |
| *** gmmaha has joined #openstack-swift | 16:41 | |
| *** diazjf has joined #openstack-swift | 16:43 | |
| *** zaitcev has joined #openstack-swift | 16:44 | |
| *** ChanServ sets mode: +v zaitcev | 16:44 | |
| *** diazjf1 has joined #openstack-swift | 16:48 | |
| *** diazjf has quit IRC | 16:50 | |
| *** klrmn has quit IRC | 17:09 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: EC: Avoid conflicts when ssync'ing fragments https://review.openstack.org/181407 | 17:14 |
| openstackgerrit | Alistair Coles proposed openstack/swift: Enable object server to return non-durable data https://review.openstack.org/215276 | 17:27 |
| *** SkyRocknRoll has quit IRC | 17:40 | |
| *** gyee has quit IRC | 17:51 | |
| *** aix has quit IRC | 17:58 | |
| notmyname | good morning. | 17:59 |
| * notmyname is in-and-out this week and next | 18:00 | |
| *** ChubYann has joined #openstack-swift | 18:00 | |
| janonymous_ | o/ cool, merry christmas nd hpy new year in advance :) | 18:01 |
| *** chsc has joined #openstack-swift | 18:01 | |
| notmyname | thanks | 18:03 |
| *** dslevin_ has joined #openstack-swift | 18:05 | |
| *** kevinc has joined #openstack-swift | 18:13 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 18:15 |
| *** rledisez has quit IRC | 18:15 | |
| *** rledisez has joined #openstack-swift | 18:18 | |
| *** klrmn has joined #openstack-swift | 18:19 | |
| *** dslevin__ has joined #openstack-swift | 18:20 | |
| *** dslevin_ has quit IRC | 18:20 | |
| *** gyee has joined #openstack-swift | 18:22 | |
| *** gyee has quit IRC | 18:25 | |
| *** geaaru has quit IRC | 18:27 | |
| *** kevinc has quit IRC | 18:30 | |
| *** acoles is now known as acoles_ | 18:32 | |
| *** eranrom has joined #openstack-swift | 18:32 | |
| *** dslevin__ has quit IRC | 18:36 | |
| *** rledisez has quit IRC | 18:37 | |
| *** petertr7_away is now known as petertr7 | 18:51 | |
| *** gyee has joined #openstack-swift | 18:53 | |
| *** dslevin_ has joined #openstack-swift | 18:54 | |
| *** trifon has joined #openstack-swift | 18:56 | |
| *** petertr7 is now known as petertr7_away | 19:02 | |
| *** changbl has joined #openstack-swift | 19:05 | |
| *** dslevin_ has quit IRC | 19:16 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: decouple versioned writes from COPY https://review.openstack.org/260179 | 19:30 |
| tdasilva | wbhuber, jrichli: ^^^ this should be the first patch in getting copy middleware fixed up | 19:31 |
| jrichli | tdasilva: thanks a bunch! I will check it out. | 19:32 |
| tdasilva | wbhuber, jrichli: tomorrow I will work on rebase copy middleware to depend on this patch | 19:32 |
| *** zhill has joined #openstack-swift | 19:34 | |
| jrichli | sounds great. I hope you get to enjoy the holidays too ;-) | 19:34 |
| tdasilva | jrichli: i will be going on break after wed! hope you have some time off too | 19:37 |
| jrichli | tdasilva: ah, that's good. I have been on break since Wednesday. Just been doing a little work here and there. I hope to get a lot accomplished between the 27th and the 4th! | 19:41 |
| *** dslevin_ has joined #openstack-swift | 19:48 | |
| *** esker has joined #openstack-swift | 19:53 | |
| *** esker has quit IRC | 19:54 | |
| *** esker has joined #openstack-swift | 19:54 | |
| *** dslevin_ has quit IRC | 20:00 | |
| *** pc6411 is now known as pchng | 20:14 | |
| openstackgerrit | Eran Rom proposed openstack/swift: Container-Sync to check the right client exception https://review.openstack.org/260204 | 20:15 |
| Zyric | Good Morning | 20:15 |
| *** zhill has quit IRC | 20:18 | |
| *** janonymous_ has quit IRC | 20:26 | |
| *** zhill has joined #openstack-swift | 20:29 | |
| mattoliverau | morning | 20:31 |
| *** zhill has quit IRC | 20:31 | |
| *** esker has quit IRC | 20:42 | |
| *** esker has joined #openstack-swift | 20:43 | |
| *** changbl has quit IRC | 21:14 | |
| *** petertr7_away is now known as petertr7 | 21:19 | |
| notmyname | hello mattoliverau and Zyric | 21:27 |
| *** asettle-afk has quit IRC | 21:37 | |
| notmyname | tdasilva: if you get a chance before you take off, could you take another glance at patch 241978 ? | 21:49 |
| patchbot | notmyname: https://review.openstack.org/#/c/241978/ - Add support for storage policies to have more than... | 21:49 |
| onovy | someone for review pls? https://review.openstack.org/#/c/251151/2 | 22:03 |
| *** petertr7 is now known as petertr7_away | 22:05 | |
| *** tonyb has joined #openstack-swift | 22:19 | |
| tonyb | notmyname: anychance you can remove your -2 from https://review.openstack.org/#/c/174343/ and https://review.openstack.org/#/c/173982/ ? | 22:21 |
| tonyb | (both kilo requirements updates) | 22:21 |
| notmyname | tonyb: why? | 22:25 |
| tdasilva | notmyname: sure | 22:26 |
| notmyname | tdasilva: thanks | 22:26 |
| tonyb | notmyname: because it will help with swift co-installability for kilo. | 22:26 |
| onovy | aaa global requirements :) | 22:28 |
| *** asettle has joined #openstack-swift | 22:28 | |
| notmyname | tonyb: our of curiosity, how so? my understanding is that the global file is used in the gates, regardless of what the per-project requirements file is | 22:29 |
| onovy | tonyb, it's moving version only up. so swift accept NEWER versions fine. co-installability is fine | 22:29 |
| notmyname | also, what onovy said ;-) | 22:30 |
| tonyb | notmyname: the upper contrainsts file is used in liberty/mitaka not in kilo | 22:30 |
| onovy | swift will work with greenlet 0.3.1 and 0.3.2 too. so use 0.3.2 from other projects (requirements) and swift will be fine too | 22:31 |
| onovy | (example) | 22:31 |
| tonyb | notmyname: the *requirements files in each project still need to be in sync or you end up with projectX installing a library that breaks projectY | 22:31 |
| onovy | no, that's package maintaner work | 22:32 |
| onovy | and i'm packing swift for debian, so i should know it :) | 22:32 |
| tonyb | onovy: but the dependancy parser in pip isn't as good as it could be (that's being worked on) so that doesn't work. | 22:32 |
| onovy | so fix pip is correct way? :) | 22:32 |
| tdasilva | Zyric: hi, re patch 241978, I just had a question about the new methods in StoragePolicyCollection, couldn't find how they are being used. what am I missing? | 22:33 |
| tonyb | onovy: like I said that's being worked on but it's complex the global-requirements is the way OpenStack as a whole deals with it. | 22:33 |
| patchbot | tdasilva: https://review.openstack.org/#/c/241978/ - Add support for storage policies to have more than... | 22:33 |
| onovy | if project X want eventlet >=0.3.2 and projectY want eventlet >=0.3.1 then 0.3.2 will be installed. -> it's working correctly | 22:33 |
| *** asettle has quit IRC | 22:34 | |
| torgomatic_ | Zyric: ping | 22:35 |
| Zyric | tdasilva: I don't believe they are used yet, it's a new feature so I just included all the functions I believed would be useful. | 22:35 |
| onovy | and i'm merging glob-req in my project (swauth). but i understand why swift is not doing the same | 22:35 |
| Zyric | torgomatic_: hello hello! Hope you're feeling better :) | 22:36 |
| torgomatic_ | Zyric: yes, for the moment :) | 22:36 |
| onovy | if swift don't need new version, why force to update? there are many envs where swift live without others openstack projects (and it our company is on of it). | 22:36 |
| torgomatic_ | Zyric: notmyname said you wanted to talk to me about the auditor stuff | 22:36 |
| Zyric | torgomatic_: That's right, I had a few questions relating to how and where it plugs into the middleware | 22:37 |
| Zyric | Which configs to edit, etc. | 22:37 |
| torgomatic_ | yeah, I should probably come up with a sample project and stick it in a review comment or something | 22:37 |
| Zyric | torgomatic_ that would be super useful. I'm having a bit of trouble getting it going in my swift system but it works fine in the tests | 22:38 |
| onovy | tonyb, so, can you explain better why it's needed? | 22:38 |
| onovy | ok, time to go to bed. good night | 22:41 |
| onovy | tonyb, continue with notmyname, thanks :) | 22:41 |
| tonyb | onovy: So you're right that the swift change only raises the minimums but the swiftclient one adds a ban on requests 2.8.0 which is needed. | 22:42 |
| notmyname | onovy: I was loving the questions you were asking :-) | 22:42 |
| tonyb | onovy: ok, sorry I've been tryign to get my thoughts in order | 22:42 |
| tonyb | notmyname, onovy: so there are 2 aspects the technical one where the python-swiftclient requirement on requests can break other openstack libraries if installed globally, and social contract that OpenStack projects opt into the requirements updates. I understadn that swift more than pretty much anything else walks the line between installing as part of OpenStack and isoloated. | 22:45 |
| tonyb | notmyname, onovy: If you can't/wont take the python-swiftclient update will you at least take a specific review for the requests change? | 22:46 |
| tonyb | notmyname, onovy: it'd be my preference to take the whoel things as it makes spotting things easier but somethign is better than notthing. | 22:47 |
| notmyname | tonyb: ...also composing my thoughts | 22:50 |
| *** dslevin_ has joined #openstack-swift | 22:51 | |
| *** dslevin_ has quit IRC | 22:55 | |
| notmyname | tonyb: I'm more open to accepting just updating the requests line | 22:57 |
| notmyname | however... | 22:57 |
| notmyname | tonyb: so I install swiftclient first and for some reason I get 2.4.0 installed. and swiftclient works. then I install something else and it upgrades requests. and swiftclient and the other thing work. or I install Other first and get requests per g-r. then I install swiftclient and it works | 22:57 |
| notmyname | so what is the problem with coinstallability? | 22:58 |
| *** km__ has joined #openstack-swift | 23:00 | |
| notmyname | tonyb: perhaps I'm missing somethign about how things are installed and the way dependencies are resolved | 23:01 |
| tonyb | notmyname: Install projectX which has !=2.4.0 , <2.8.0 and that's fine. You install swiftclient which just says > 1.x (or similar) and get 2.8.0 which *may* be fine for swift or not (I'd need to check why we banned 2.8.0) but will break projectX | 23:02 |
| tonyb | but that may only be true if you use -U | 23:02 |
| tonyb | notmyname: I have to admit I've confused myself now. | 23:02 |
| *** dslevin_ has joined #openstack-swift | 23:03 | |
| notmyname | tonyb: to be clear, in general, I really like and support the idea behind global requirements. we should make things easier for packagers (including distros). I'm all for that, and I don't want to stand in the way | 23:03 |
| notmyname | which is why I'd be more open to just the requests line | 23:03 |
| notmyname | especially if anyone can tell me why things break as-is today | 23:04 |
| *** chsc has quit IRC | 23:04 | |
| notmyname | part of making things easier for packagers, IMO, is not upgrading packages every time someone starts using a new feature | 23:05 |
| tonyb | notmyname: Yeah we're working on how to do that better but sadly we have legacy we need to carry around. | 23:06 |
| notmyname | but what is that? what is broken today that accepting this patch (or a more limited requests-only one) will fix? | 23:06 |
| tonyb | notmyname: so I think the swiftclient requirements is only a problem iff someone pip install -U's swftclient | 23:06 |
| notmyname | which is effectively what's done if you install without -U and don't have requests already installed | 23:07 |
| lifeless | ^ that | 23:08 |
| lifeless | notmyname: so that would be a good and interesting discussion to have (not upgrading minimums rapidly) | 23:08 |
| lifeless | notmyname: and its one I'd like to have, but its (hopefully obviously) non-trivial | 23:09 |
| notmyname | so to make sure I understand, the problem needing fixing is the situation where swiftclient is installed before other openstack projects and also is the first app on the system to pull in requests and then you want to bring in other openstack projects who have merged the requests limiter from g-r? | 23:11 |
| notmyname | lifeless: I too would like to have it (ie see my review comments from 2014 on this topic on the proposal bot patches). but I'm not sure it's obviously non-trivial. at least not to those of us who don't have a deep history of distro building | 23:13 |
| tonyb | notmyname: Yes. | 23:14 |
| tonyb | lifeless: I'd like to be in the room while that's discussed. | 23:15 |
| notmyname | tonyb: is the implication that you'd also need a kilo release and the same patches to be in master (and what about liberty)? | 23:15 |
| *** Jeffrey4l has joined #openstack-swift | 23:15 | |
| lifeless | notmyname: so the non-triviality is that we have no described mechanism to express 'using requests < X will be horribly insecure' | 23:17 |
| lifeless | notmyname: *other* than setting a minimum | 23:17 |
| lifeless | (substitute any other thing there too - 'use Cinder 2015.4' | 23:18 |
| notmyname | lifeless: you've just introduced a new variable. I though we were talking about a solver for versions for generic dependency X. is there some other technical thing about requests? | 23:18 |
| lifeless | notmyname: no, its not a solver issue; its that making things within a code base only work when a dependency is above some version - higher than the expressed minimum in the dependency list - creates a sort of mystery meat experience for users | 23:19 |
| tonyb | notmyname: to be frank I haven't looked at liberty becase I decided to start with kilo | 23:19 |
| tonyb | notmyname: I don't think this mandates a release but I'd need to think harder about that. | 23:19 |
| notmyname | tonyb: is there a specific security thing in requests that necessitates a min version bump, as lifeless hints? | 23:20 |
| tonyb | notmyname: I'll look. | 23:21 |
| lifeless | notmyname: sorry, I was using a single example of the reason why 'not raising minimums' is non-trivial. | 23:21 |
| lifeless | notmyname: for requests *in specific*, let me check the reasons we've been updating minimums | 23:21 |
| lifeless | so 2.8.0 broke some gate jobs | 23:23 |
| lifeless | bug 1503768 | 23:23 |
| openstack | bug 1503768 in python-glanceclient "keystoneclient.exceptions.RequestTimeout with requests==2.8.0" [Undecided,New] https://launchpad.net/bugs/1503768 - Assigned to Flavio Percoco (flaper87) | 23:23 |
| notmyname | doesn't this effectively turn in to "you must be running the latest release of any dependency" if any project hitting any bug in a dependency requires all openstack projects to raise minimum versions to the first version after the bug int he dependency is fixed? | 23:25 |
| tonyb | notmyname: it wasn't a security issue that raised the minimum. It was oslo.vmware needed to use file:// urls which (aparently) weren't supported in <2.2.0 | 23:25 |
| pdardeau | notmyname: | 23:26 |
| pdardeau | notmyname: ping | 23:26 |
| tonyb | lifeless: yeah that's what capped the requests | 23:26 |
| lifeless | we bumped to 2.6.0 because of issues with docker-py | 23:26 |
| notmyname | pdardeau: what's up? | 23:27 |
| lifeless | 2.2.0 was because of oslo.vmware file:// support | 23:27 |
| tonyb | https://review.openstack.org/#/c/123714/ is what raised kilo from requests>=1.2.1 to requests>=2.2.0 | 23:27 |
| pdardeau | notmyname: hey, just wanted to give an update on the device count increase feature | 23:27 |
| lifeless | however | 23:28 |
| lifeless | http://www.cve.mitre.org/cgi-bin/cvename.cgi?name=2015-2296 | 23:28 |
| pdardeau | i'm planning to push wip next week and possibly even the non-wip by end of next week | 23:28 |
| notmyname | pdardeau: nice | 23:28 |
| lifeless | requests 2.6.0 is the minimum to avoid known security issues | 23:28 |
| lifeless | ( from http://docs.python-requests.org/en/latest/community/vulnerabilities/ ) | 23:28 |
| pdardeau | notmyname: i'll be off on vacation remainder of this week, but will be working all next week | 23:28 |
| notmyname | pdardeau: me too ;-) | 23:28 |
| lifeless | and 2.3.0 has a couple of bad vulns | 23:29 |
| lifeless | notmyname: so the interaction of the solver here is that only the first reference to a package is honoured by pip | 23:29 |
| pdardeau | all: best wishes for the holidays and i'll return next Monday | 23:29 |
| lifeless | notmyname: we're working on that, and once fixed we can consider doing partial-upgrades of dependencies across openstack | 23:29 |
| notmyname | pdardeau: thanks for the update. I'm off this week and next and just in and out. I'm looking forward to it! | 23:29 |
| notmyname | the security issue(s) is very different than file or docker support. that's a good reason to upgrade the min version | 23:30 |
| lifeless | notmyname: one of the complexities of embracing skewed minimums across openstack though is testing: and note that *all* distributors distribute only a single version of requests. | 23:31 |
| lifeless | notmyname: so we don't help distributors at all by having differing minimums: they upgrade as soon as anything they distribute needs the upgrade | 23:31 |
| notmyname | tonyb: so if you submit a patch for the requests line and leave a comment (or commit message) referencing the security stuff to stable/kilo for swiftclient, I'll approve it | 23:31 |
| *** trifon has quit IRC | 23:33 | |
| notmyname | lifeless: I understand that. but this isn't a problem that came into being with openstack. it's something that distros have always been doing. one compatible version number is very different than your example of a project enabling/disabling things based on newer versions | 23:34 |
| notmyname | *newer versions of dependencies. and I include "having security holes" in that too | 23:35 |
| *** ianbrown_ has quit IRC | 23:35 | |
| notmyname | and having a reasonable set of "here's requirements" common across openstack is good. like I said, I agree with and support the ideas behind global requirements | 23:35 |
| *** ianbrown_ has joined #openstack-swift | 23:35 | |
| notmyname | I just have some problems with the practical aspects of how it's implemented | 23:36 |
| *** kei_yama has joined #openstack-swift | 23:36 | |
| *** trifon has joined #openstack-swift | 23:37 | |
| lifeless | notmyname: so, we need to get those laid out somewhere; I can't come up with creative solutions until I know the concerns :) | 23:40 |
| notmyname | 1) I don't like that any other project using a just-released feature in a third party library can compel a single project's deployers to rebuild packages when they upgrade (for changes that project doesn't require) | 23:41 |
| notmyname | 2) I don't like the exact string match solver for g-r. it means I can't just add an exclusion or satisfy requirements by accepting things more broadly (eg libfoo>=1.0 in requirements is not compatible with libfoo>=1.5 in g-r) | 23:41 |
| notmyname | 3) g-r assumes all requirements are found on pypi and installable with pip. ie python libraries. that's not the case at all, and thus the requirements file is always incomplete and doesn't cover all the dependency concerns | 23:41 |
| lifeless | notmyname: the reason I talked about enabling/disabling things is that there are really three implementation choices: all the same, different everywhere, and all-the-same-oldest-version-with-internal-enablement-of-new-things | 23:41 |
| lifeless | notmyname: so 3) - have you looked at bindep ? | 23:43 |
| notmyname | I don't know what that is | 23:43 |
| lifeless | notmyname: https://github.com/openstack-infra/bindep/blob/master/README.rst | 23:43 |
| *** ho has quit IRC | 23:44 | |
| lifeless | notmyname: its a thing I wrote to solve the inability to refer to system packages as dependencies | 23:44 |
| lifeless | notmyname: so I'd like to see good answers for 3, its a non-tiny field itself: do we have one project coordinating all the deps, | 23:45 |
| lifeless | or do we have one coordination point per language/packaging system ? | 23:45 |
| lifeless | notmyname: on 1 and 2, they are the result of the current choice from the solutioon space I referred to above | 23:46 |
| notmyname | I'm pretty sure one of the main reasons docker et al get so much hype is because devs can package whatever latest greatest they want in their sandbox without the sysadmins having to make a dependnecy solver ;-) | 23:46 |
| lifeless | notmyname: the choices again: all-the-same-highest-minimum, all-the-same-lowest-minimum-plus-selective-enablement, project-local-minimums | 23:48 |
| lifeless | notmyname: currently openstack's official policy is all-the-same-highest-minimum. | 23:48 |
| lifeless | notmyname: it causes 1) above | 23:49 |
| lifeless | notmyname: and 2) above | 23:49 |
| notmyname | I understand the logic that gets there. I just don't like the destination | 23:50 |
| lifeless | notmyname: if we adopt project-local-minimums | 23:50 |
| lifeless | notmyname: we need to have a solver in pip first (to avoid https://github.com/pypa/pip/issues/988 ) | 23:52 |
| lifeless | notmyname: we also need some rule to ensure that there is *a* consistent version across all of openstack [or we can't co-install everything] | 23:52 |
| lifeless | notmyname: we also have a testing gap that can arise - right now we don't test minimums at all, and if we do we can at least test one minimum across everything | 23:53 |
| lifeless | if there are project-local minimums | 23:53 |
| lifeless | we have to figure out some way to test each project with its actual minimums | 23:54 |
| lifeless | notmyname: sure, I'm trying to figure out some way to meet your needs | 23:56 |
| *** changbl has joined #openstack-swift | 23:57 | |
| lifeless | anyhoo | 23:58 |
| *** ho has joined #openstack-swift | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!