| *** catintheroof has quit IRC | 00:29 | |
| *** catintheroof has joined #openstack-manila | 00:30 | |
| *** catintheroof has quit IRC | 00:30 | |
| *** tuanluong has joined #openstack-manila | 00:41 | |
| *** catintheroof has joined #openstack-manila | 00:42 | |
| *** gcb has quit IRC | 00:44 | |
| *** cfey has quit IRC | 01:15 | |
| *** cfey has joined #openstack-manila | 01:16 | |
| *** gouthamr has quit IRC | 01:48 | |
| *** gouthamr has joined #openstack-manila | 01:49 | |
| bswartz | vkmc: sorry for late resonse | 01:56 |
|---|---|---|
| bswartz | vkmc: I did that review not long before I left the office | 01:56 |
| bswartz | vkmc: I see no reason that we can't make every job both ipv4 and ipv6 aware, and run both kinds of tests in the same job | 01:57 |
| bswartz | there's no reason to create more jobs to do that -- we just need to make existing job ipv6-aware and make sure it's possible to disable ipv6-related tests on plaforms where they won't work | 01:58 |
| *** kaisers_ has joined #openstack-manila | 02:17 | |
| *** kaisers_ has quit IRC | 02:22 | |
| *** wangqun has joined #openstack-manila | 03:16 | |
| *** wangqun has quit IRC | 03:17 | |
| *** wangqun has joined #openstack-manila | 03:18 | |
| *** catintheroof has quit IRC | 03:57 | |
| *** catintheroof has joined #openstack-manila | 03:58 | |
| *** catintheroof has quit IRC | 03:59 | |
| *** gcb has joined #openstack-manila | 04:19 | |
| *** hhgregg has joined #openstack-manila | 04:23 | |
| *** hhgregg has left #openstack-manila | 04:23 | |
| *** kaisers_ has joined #openstack-manila | 04:33 | |
| *** kaisers_ has quit IRC | 04:38 | |
| *** wangqun has quit IRC | 04:45 | |
| *** wangqun has joined #openstack-manila | 04:46 | |
| *** kaisers_ has joined #openstack-manila | 05:23 | |
| *** kaisers_ has quit IRC | 05:25 | |
| *** gouthamr has quit IRC | 05:29 | |
| *** sandanar has joined #openstack-manila | 05:48 | |
| *** mtanino has joined #openstack-manila | 05:57 | |
| *** markstur has quit IRC | 06:21 | |
| *** markstur has joined #openstack-manila | 06:22 | |
| *** markstur has quit IRC | 06:26 | |
| *** nkrinner_afk is now known as nkrinner | 06:37 | |
| *** a-pugachev has joined #openstack-manila | 07:19 | |
| *** kaisers_ has joined #openstack-manila | 07:26 | |
| *** kaisers_ has quit IRC | 07:30 | |
| *** jprovazn has joined #openstack-manila | 07:32 | |
| *** a-pugachev has quit IRC | 07:55 | |
| *** lpetrut has joined #openstack-manila | 08:03 | |
| *** lpetrut has quit IRC | 08:12 | |
| openstackgerrit | Peter Wang proposed openstack/manila master: Handle ssl for VNX manila driver https://review.openstack.org/440364 | 08:45 |
| *** sandanar has quit IRC | 09:17 | |
| *** tinwood is now known as tinwood_swap | 09:18 | |
| *** markstur has joined #openstack-manila | 09:22 | |
| *** markstur has quit IRC | 09:26 | |
| *** swapnil-nilangek has joined #openstack-manila | 09:30 | |
| *** a-pugachev has joined #openstack-manila | 09:41 | |
| *** dsariel has quit IRC | 09:57 | |
| *** wangqun has quit IRC | 09:58 | |
| *** a-pugachev has quit IRC | 10:02 | |
| *** kaisers_ has joined #openstack-manila | 10:23 | |
| *** kaisers_ has quit IRC | 10:23 | |
| *** kaisers_ has joined #openstack-manila | 10:24 | |
| *** a-pugachev has joined #openstack-manila | 10:31 | |
| *** openstackgerrit has quit IRC | 10:33 | |
| *** kaisers_ has quit IRC | 10:38 | |
| *** ganso has joined #openstack-manila | 10:49 | |
| *** openstackgerrit has joined #openstack-manila | 10:51 | |
| openstackgerrit | Jan Provaznik proposed openstack/manila master: Provide filter name in user messages https://review.openstack.org/443102 | 10:51 |
| *** markstur has joined #openstack-manila | 11:10 | |
| *** markstur has quit IRC | 11:15 | |
| *** a-pugachev has quit IRC | 11:17 | |
| *** a-pugachev has joined #openstack-manila | 11:20 | |
| *** jprovazn has quit IRC | 11:28 | |
| *** swapnil-nilangek has quit IRC | 11:45 | |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/puppet-manila master: rabbitmq SSL: check puppet resource instead of actual config https://review.openstack.org/444244 | 11:54 |
| openstackgerrit | Juan Antonio Osorio Robles proposed openstack/puppet-manila master: rabbitmq SSL: check puppet resource instead of actual config https://review.openstack.org/444244 | 12:14 |
| *** rraja has joined #openstack-manila | 12:24 | |
| *** catintheroof has joined #openstack-manila | 12:34 | |
| *** kaisers_ has joined #openstack-manila | 12:38 | |
| *** kaisers_ has quit IRC | 12:43 | |
| *** databus23_ has joined #openstack-manila | 12:53 | |
| *** dsariel has joined #openstack-manila | 12:54 | |
| *** jprovazn has joined #openstack-manila | 12:57 | |
| arnewiebalck | There are no per share type quota classes in Manila? | 12:59 |
| *** markstur has joined #openstack-manila | 12:59 | |
| arnewiebalck | Use case: We have multiple backends accessible via different share types and would like to control the usage via the quota the user have per type. | 13:00 |
| *** markstur has quit IRC | 13:04 | |
| *** gcb has quit IRC | 13:08 | |
| *** rraja has quit IRC | 13:11 | |
| *** rraja has joined #openstack-manila | 13:11 | |
| *** lpetrut has joined #openstack-manila | 13:19 | |
| *** kaisers_ has joined #openstack-manila | 13:27 | |
| arnewiebalck | I guess type-access-add is meant to cover this use case? | 13:37 |
| tbarron | arnewiebalck: I think no one objected to porting per-share-type quotas over from cinder's per-volume-type quotas, just hasn't been done yet. | 13:43 |
| *** kaisers_ has quit IRC | 13:46 | |
| arnewiebalck | tbarron: ok, thx. Should I open a bug for this? | 13:49 |
| tbarron | arnewiebalck: arguably it's not a bug, would be a blueprint for new functionality | 13:50 |
| tbarron | bswartz: ^^^ ? | 13:50 |
| *** dustins has joined #openstack-manila | 13:57 | |
| *** tuanluong has quit IRC | 14:01 | |
| *** xyang1 has joined #openstack-manila | 14:12 | |
| *** porrua has joined #openstack-manila | 14:13 | |
| bswartz | def not a bug | 14:15 |
| *** nkrinner is now known as nkrinner_afk | 14:17 | |
| arnewiebalck | bswartz: a bp then? | 14:17 |
| bswartz | arnewiebalck: maybe even a brief spec | 14:19 |
| bswartz | if we can hash out any design issues in a spec review the code review will go faster | 14:19 |
| bswartz | however if the code will be pretty trivial then I would consider it unneceesary | 14:20 |
| *** kaisers_ has joined #openstack-manila | 14:20 | |
| * bswartz spelling fail | 14:20 | |
| bswartz | unnecessary | 14:20 |
| *** kaisers_ has quit IRC | 14:25 | |
| *** gouthamr has joined #openstack-manila | 14:25 | |
| *** tommylikehu_ has joined #openstack-manila | 14:26 | |
| arnewiebalck | bswartz: I didn’t look so can’t judge atm how complicated the code would be. | 14:30 |
| *** rraja has quit IRC | 14:30 | |
| arnewiebalck | bswartz: From our experience with Cinder, though, I’d think controlling the resource provisioning via type-based quotas would be useful. | 14:32 |
| arnewiebalck | bswartz: And I guess noone would trust my judgement if I came back in a couple days saying “this is trivial“ ;) | 14:34 |
| arnewiebalck | bswartz: I’m happy to file a brief spec, then we take it from there. Sounds ok? | 14:36 |
| *** cknight has joined #openstack-manila | 14:43 | |
| bswartz | arnewiebalck: yeah that sounds great | 14:44 |
| bswartz | arnewiebalck: is there a cinder spec we can copy from? or did cinder do this before they started doing specs? | 14:44 |
| bswartz | vkmc: ping me when you're around, if you still have questions about my thinking around ipv6 tests | 14:45 |
| *** kaisers_ has joined #openstack-manila | 14:46 | |
| *** markstur has joined #openstack-manila | 14:48 | |
| tbarron | bswartz: arnewiebalck This was prior to cinder specs process: https://review.openstack.org/#/c/25059/ | 14:51 |
| arnewiebalck | bswartz: tbarron: for the spec, this is the closest I found: https://github.com/openstack/cinder-specs/blob/eee5dde83aebfc94ad9c67550a8337e175237665/specs/kilo/support-volume-backup-quota.rst | 14:52 |
| *** markstur has quit IRC | 14:53 | |
| vkmc | bswartz, hey there! thanks for your reply yesterday | 14:54 |
| vkmc | I left a comment on Gerrit for the patch set as well | 14:55 |
| *** kaisers_ has quit IRC | 14:55 | |
| vkmc | I'm still doubtful on how we are supposed to make that work, considering the fact that when we deploy the backends you can either use one IP version protocol or the other | 14:56 |
| bswartz | vkmc: in that case step one is fixing that limitation so that we can use both at the same time | 14:56 |
| arnewiebalck | bswartz: tbarron: I’ll prepare a spec based on the one I found and the commit tbarron found … from the Cinder patch I wouldn’t say it’s completely trivial. | 14:56 |
| vkmc | e.g. in LVM you have the option MANILA_LVM_SHARE_EXPORT_IP that takes the ip you want to use | 14:56 |
| vkmc | for the export location | 14:57 |
| bswartz | yes i know | 14:57 |
| vkmc | bswartz, that would require changing the drivers to adjust that | 14:57 |
| bswartz | let me see if that option might accept a list | 14:57 |
| *** rraja has joined #openstack-manila | 14:57 | |
| bswartz | nope | 14:58 |
| bswartz | shouldn't be too hard to change that option from a StrOpt to a ListOpt though | 14:58 |
| bswartz | now that I think about it we'd want to rename it though | 14:58 |
| bswartz | lvm_share_export_ip -> lvm_share_export_ips | 14:58 |
| vkmc | bswartz, and... logic-wise, which one you would pick? | 14:59 |
| bswartz | vkmc: do you want to work on that or should I? | 14:59 |
| vkmc | if you have a dual stack then default to ipv6? | 14:59 |
| tbarron | bswartz: vkmc so I'm all for that but in my mind that was a step 2. Step 1 was getting the jobs in so that we could prove that the image works with IPv6. | 14:59 |
| bswartz | tbarron: this proposal would also allow dual IPv4 addresses, which could be a useful thing | 14:59 |
| tbarron | yeah I'm for it, it just wasnt' what I asked vkmc to sign up for | 15:00 |
| *** catintheroof has quit IRC | 15:00 | |
| bswartz | but it would allow us to start setting up jobs with IPv4+IPv6 too | 15:00 |
| bswartz | let's get our 3 heads together and split up this work in a way that makes sense | 15:00 |
| bswartz | I'm happy do any LVM work that needs doing | 15:01 |
| bswartz | or changes to tests | 15:01 |
| *** catintheroof has joined #openstack-manila | 15:01 | |
| bswartz | tbarron vkmc: do we already have a basic LVM job working on Centos? | 15:01 |
| vkmc | bswartz, we do | 15:02 |
| bswartz | vkmc: and what exactly was the goal of your next change? | 15:02 |
| vkmc | bswartz, next step was adding the lvm configured with an ipv6 export location | 15:02 |
| bswartz | okay so I'd like to do that without creating any new jobs | 15:02 |
| bswartz | the best way I can think of is to make existing jobs able to deal with multiple IPs | 15:03 |
| vkmc | to test the ipv6 driver | 15:03 |
| vkmc | as per discussion in the ptg <- https://etherpad.openstack.org/p/manila-pike-ptg-wednesday | 15:03 |
| bswartz | do you see any problems with my proposal of simply allowing multiple IPs in the LVM driver? | 15:03 |
| vkmc | I'm not aware of how much work is needed to actually make the lvm driver able to mange multiple IPs | 15:04 |
| tbarron | we were just proceeding incrementally: 1) ipv4 job for lvm with CentOS (check, it works). (2) ipv6 lvm job with CentOS (we think it will work), (3) then work with you to consolidate jobs, modify lvm driver for multiple exports, etc. | 15:05 |
| vkmc | do you think is something we can do in the short term? | 15:05 |
| tbarron | I don't necessarily object to squashing #2 and #3, just am indicating what we were thinking | 15:05 |
| * tbarron is a methodical turtle | 15:05 | |
| bswartz | tbarron: but how would (2) ever work without tommy's code merged? | 15:06 |
| tbarron | for lvm you just have to set manila.conf lvm_export_address config (or whatever it's called) | 15:07 |
| tbarron | currently it's set to IPv4 address | 15:07 |
| bswartz | lvm_share_export_ip | 15:07 |
| bswartz | no we've proven that that blows up | 15:07 |
| bswartz | the LVM driver needs changes to work with ipv6 at all | 15:07 |
| * bswartz goes looking for tommy's patch | 15:08 | |
| tbarron | ah, I was forgetting what those I guess. | 15:08 |
| tbarron | what those were | 15:08 |
| *** jprovazn has quit IRC | 15:09 | |
| vkmc | hmm all right | 15:09 |
| vkmc | so... I should abandon that change | 15:09 |
| vkmc | and actually look on how we can make LVM work with a dual stack | 15:09 |
| tbarron | vkmc: well, wait a minute :D | 15:09 |
| vkmc | sure | 15:09 |
| bswartz | tbarron: https://review.openstack.org/#/c/406776/33/manila/share/drivers/lvm.py | 15:09 |
| tbarron | oh yeah, that's pretty trivial though. | 15:10 |
| arnewiebalck | Is there a way to change the visibility of a share type after creation? | 15:10 |
| arnewiebalck | Something like “type-update”. | 15:11 |
| tbarron | bswartz: but your point is valid, we'll have to have that change too to make it work. | 15:11 |
| bswartz | tbarron: the most incremental baby steps I can think of is to make all jobs IPv4+IPv6 capable and to disable IPv6 by default in all jobs | 15:12 |
| tbarron | bswartz: theoretically, vkmc could do #2, we could do this change of tommy's and see it work, and then we could do a change for multiple export locations. | 15:12 |
| *** mtanino has quit IRC | 15:12 | |
| bswartz | then as we start to tinker with patches that enable IPv6 we enable IPv6 in the CentOS job only | 15:12 |
| *** rraja has quit IRC | 15:12 | |
| bswartz | the main downside is that I'll have to change my own dev/test workflow | 15:13 |
| *** eharney has joined #openstack-manila | 15:13 | |
| bswartz | because I'm not setup to develop or test on CentOS | 15:13 |
| tbarron | bswartz: I'm not dugin on incrementalism or whatever. | 15:14 |
| tbarron | just wanted to see lvm with ipv6 with CentOS image working, then tackle the bigger issues. | 15:14 |
| bswartz | tbarron: we have a dependency loop we need to break though | 15:15 |
| *** chlong_ has joined #openstack-manila | 15:15 | |
| bswartz | the tests will depend on the code and the code will depend on the tests | 15:15 |
| tbarron | we can put in a nonvoting job that will fail till we get new lvm code in place | 15:15 |
| bswartz | O dpm | 15:16 |
| tbarron | ? | 15:16 |
| bswartz | I don't think you'll be able to do that | 15:16 |
| bswartz | sorry my fat fingers can't type | 15:16 |
| tbarron | vkmc: does your latest patch actually gate on the lvm job passing? | 15:17 |
| bswartz | tbarron: a bunch of patches need to be tested together to get a working IPv6 result | 15:18 |
| bswartz | tbarron: one option is just to start merging the patches without test coverage (assuming they don't break any existing stuff) | 15:18 |
| bswartz | and plan to get the test coverage sorted out in the next 4 months | 15:18 |
| vkmc | tbarron, what do you mean? | 15:19 |
| bswartz | tbarron: patches to project-config don't actually run the manila tests so you can't see if they work or not | 15:19 |
| tbarron | vkmc: that's what I mean ^^^ | 15:20 |
| vkmc | tbarron, ^ | 15:20 |
| vkmc | that | 15:20 |
| vkmc | yes | 15:20 |
| vkmc | we don't know yet | 15:20 |
| tbarron | so it could go in & it will fail when run in manila itself until we merge this code: https://review.openstack.org/#/c/406776/33/manila/share/drivers/lvm.py | 15:21 |
| vkmc | we would need to run tests on that environment built manually if we want to check before hand if tests are working for that set of configurations | 15:21 |
| bswartz | tbarron: that's my expectation | 15:22 |
| tbarron | so as long as vkmc's ipv6 job doesn't vote it won't break anything and we can try tommy's code against it | 15:22 |
| bswartz | tbarron: but my higher level point is that we can do everything we need to without further changes to project-config | 15:22 |
| tbarron | bswartz: not objecting, just wasn't what I was thinking. | 15:23 |
| bswartz | changes to project-config are annoying because they're slow (must wait for infra to merge) and not testable before they merge | 15:23 |
| bswartz | all the more reason not to make "temporary" changes which we know we'll need to back out later | 15:24 |
| tbarron | so we use the existing ipv4 job then and modify manila/devstack/plugin and lvm ? | 15:24 |
| tbarron | existing ipv4 centOS job that is | 15:25 |
| vkmc | I need to step out for a moment, brb | 15:25 |
| tbarron | kk | 15:25 |
| bswartz | tbarron: I'm already working on the LVM driver multi-IP change | 15:26 |
| tbarron | bswartz: sounds good :D | 15:26 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila-image-elements master: Allow built image be available via local webserver https://review.openstack.org/444353 | 15:28 |
| *** markstur has joined #openstack-manila | 15:30 | |
| bswartz | tbarron: uh oh this is gonna be ugly | 15:31 |
| bswartz | it touches the helper code that's shared with generic | 15:31 |
| tbarron | yuck | 15:32 |
| bswartz | tbarron: I'll try to cover up the bad taste | 15:33 |
| *** dsariel has quit IRC | 15:34 | |
| tbarron | arnewiebalck: I don't know a way to change a share-type from is-public to private or vice-versa after it is created. | 15:54 |
| tbarron | anyone know any fundamental reason not to enable this toggle? ^^^^ | 15:55 |
| tbarron | Or a way to do it that I'm missing? | 15:55 |
| arnewiebalck | tbarron: Cinder offers the “type-update” subcommand which can change the visibility, the name and the descr of a volume. | 15:57 |
| *** eharney has quit IRC | 16:03 | |
| *** tommylikehu_ has quit IRC | 16:07 | |
| tbarron | arnewiebalck: https://review.openstack.org/#/c/206788/ | 16:08 |
| arnewiebalck | tbarron: Thanks! Doesn’t look too complicated … to be referred to in another spec? ;) | 16:10 |
| tbarron | arnewiebalck: yeah, I think this is rinse and repeat | 16:11 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila-image-elements master: Allow built image be available via local webserver https://review.openstack.org/444353 | 16:11 |
| openstackgerrit | Merged openstack/puppet-manila master: rabbitmq SSL: check puppet resource instead of actual config https://review.openstack.org/444244 | 16:14 |
| *** eharney has joined #openstack-manila | 16:17 | |
| *** chlong_ has quit IRC | 16:19 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: [CI] Add support for CI jobs with custom images https://review.openstack.org/444406 | 16:48 |
| *** logan_ has joined #openstack-manila | 16:56 | |
| *** logan_ is now known as Guest29589 | 16:57 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: [CI] Add support for CI jobs with custom images https://review.openstack.org/444406 | 16:58 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila-image-elements master: Allow built image be available via local webserver https://review.openstack.org/444353 | 16:58 |
| *** logan- has quit IRC | 16:59 | |
| *** Guest29589 is now known as logan- | 17:03 | |
| *** kaisers_ has joined #openstack-manila | 17:07 | |
| *** catintheroof has quit IRC | 17:07 | |
| *** catintheroof has joined #openstack-manila | 17:08 | |
| *** kaisers_ has quit IRC | 17:12 | |
| *** a-pugachev has quit IRC | 17:18 | |
| bswartz | bswartz: this change is going to be large :-( | 17:21 |
| *** a-pugachev has joined #openstack-manila | 17:21 | |
| * bswartz wonders why he's talking to himself | 17:21 | |
| bswartz | tbarron: this change is going to be large :-( | 17:22 |
| bswartz | there is tons of code in manila that assumes exactly 1 public IP per share server | 17:22 |
| tbarron | yup | 17:22 |
| tbarron | no way to divide and conquer and avoid the DHSS=True side for now? | 17:23 |
| bswartz | tbarron: I *am* doing that and it's still massive | 17:23 |
| tbarron | kk | 17:23 |
| tbarron | bswartz: kk | 17:23 |
| bswartz | the diff is up to 366 lines and I've just gotten started | 17:24 |
| bswartz | I expect pep8 to throw around 50 errors | 17:25 |
| tbarron | that's par for the course | 17:25 |
| bswartz | at least my unit test failures are down from 50 to 8 | 17:25 |
| *** JoseMello has joined #openstack-manila | 17:31 | |
| bswartz | 0 unit test failures and wait for it.... | 17:31 |
| * dustins crosses fingers | 17:32 | |
| bswartz | only 4 pep8 errors | 17:32 |
| bswartz | not bad | 17:32 |
| bswartz | on that note I think I will get lunch | 17:32 |
| tbarron | bon appetit | 17:32 |
| dustins | enjoy! | 17:33 |
| *** a-pugachev has quit IRC | 17:33 | |
| *** lpetrut has quit IRC | 17:45 | |
| vkmc | bswartz, I was afraid of that :| | 17:52 |
| vkmc | bswartz, is that only for lvm? | 17:52 |
| openstackgerrit | Tom Barron proposed openstack/manila master: Deprecate 'ganesha_nfs_export_options' https://review.openstack.org/444441 | 18:03 |
| *** sapcc-bot2 has quit IRC | 18:10 | |
| *** sapcc-bot has joined #openstack-manila | 18:10 | |
| bswartz | vkmc: you'll see soon, but as a preview, because I want to allow 2 or more IPs on the LVM driver, I'm forced to update the helpers and everything that touches those helpers, which is a lot of code | 18:14 |
| bswartz | this is all technical debt which is not strictly related to or required for IPv6, but it's a good time to fix it | 18:15 |
| *** eharney has quit IRC | 18:35 | |
| *** eharney has joined #openstack-manila | 18:51 | |
| *** eharney_ has joined #openstack-manila | 18:59 | |
| *** eharney has quit IRC | 19:01 | |
| *** eharney_ is now known as eharney | 19:02 | |
| *** kaisers_ has joined #openstack-manila | 19:08 | |
| *** kaisers_ has quit IRC | 19:12 | |
| tbarron | bswartz: vponomaryov w.r.t. deprecation of v1 API see smcginnis note on the issue in cinder: http://lists.openstack.org/pipermail/openstack-dev/2017-March/113727.html | 19:16 |
| *** markstur has quit IRC | 19:31 | |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: [WIP] Allow 2 or more export IPs for LVM driver https://review.openstack.org/444479 | 19:33 |
| bswartz | tbarron: I don't see an "issue" there, it looks like they're going to overachieve and do a complete removal | 19:35 |
| tbarron | bswartz: it was just FYI | 19:35 |
| bswartz | k | 19:35 |
| bswartz | tbarron: ^ you can see my thinking on multiple-IP support in LVM above | 19:36 |
| *** chlong_ has joined #openstack-manila | 19:42 | |
| bswartz | who understand option deprecation? | 19:46 |
| bswartz | there are now like 5 different deprecation-oriented flags | 19:46 |
| bswartz | who understands option deprecation? | 19:46 |
| tbarron | jungleboyj: ^^^ iirc you have fu in this area | 19:48 |
| bswartz | tbarron: he's supposed to be on vacation | 19:48 |
| tbarron | bswartz: oh | 19:49 |
| bswartz | tbarron: see #cinder, 12:07 PM EST | 19:49 |
| tbarron | bswartz: well if he can drink at noon and talk there I'm sure he could give sound advice here too | 19:51 |
| tbarron | on second thought, it's three hours later... | 19:51 |
| smcginnis | lol | 19:52 |
| bswartz | I drink and code all the time | 19:53 |
| tbarron | I'm going to start handing out CWIs. | 19:54 |
| tbarron | or CUIs, depending on the state one lives in. | 19:54 |
| tbarron | no pun intended on 'state' | 19:54 |
| bswartz | you just have to make sure you don't write the tests while drunk too | 19:54 |
| bswartz | ah crap | 19:56 |
| bswartz | I'm failing py35 tests | 19:56 |
| bswartz | how do I run unit tests with py35 on my dev system? | 19:57 |
| tbarron | tox -epy35 | 19:57 |
| bswartz | ty | 19:58 |
| * bswartz still uses run_tests.sh all the time | 19:58 | |
| tbarron | tox runs them now but you might need to 'rm -rf .tox .testrepository' on a pre-existing workspace | 19:58 |
| tbarron | 'tox' by itself I mean: py35, then py27, then pep8 | 19:59 |
| bswartz | ugh | 19:59 |
| bswartz | py35 is unhappy on my box | 19:59 |
| tbarron | what's it doing? | 19:59 |
| bswartz | error: command 'x86_64-linux-gnu-gcc' failed with exit status 1 | 19:59 |
| bswartz | I'm sure I'm missing a bunch of *-devel packages | 20:00 |
| tbarron | *-devel ? is this fedora or ubuntu? | 20:02 |
| bswartz | sorry -dev | 20:02 |
| tbarron | kk | 20:02 |
| bswartz | in my brain they're interchangeable | 20:02 |
| tbarron | try 'apt-get install python3-devel' | 20:03 |
| bswartz | too much exposure to Fedora | 20:03 |
| tbarron | I mean python3-dev :D | 20:03 |
| bswartz | db type could not be determined | 20:04 |
| bswartz | ERROR: InvocationError: '/opt/stack/manila/.tox/py35/bin/ostestr' | 20:04 |
| tbarron | rm -rf .testrepository | 20:04 |
| tbarron | you shouldn't ever have to do that again | 20:04 |
| bswartz | that's a neat trick | 20:04 |
| *** gouthamr has quit IRC | 20:04 | |
| *** JoseMello has quit IRC | 20:05 | |
| tbarron | And 'rm -rf .tox' for good measure | 20:05 |
| bswartz | ugh there's something evil going on | 20:10 |
| tbarron | with python3 or with your evil code? | 20:11 |
| tbarron | overall the change looks good to me btw | 20:11 |
| tbarron | though we should really be testing on more than single-element lists of addresses eventually | 20:12 |
| bswartz | unit tests that pass on 2.7 fail on 3.5 | 20:12 |
| tbarron | i'll pull the changeset and try it, that sounds intriguing | 20:12 |
| tbarron | bbiam | 20:12 |
| bswartz | tbarron: this is just me trying to un-break everything after enabling a list of IPs. obviously more test coverage is needed to actually test that lists of IPs work as expected | 20:13 |
| *** dustins has quit IRC | 20:13 | |
| *** gouthamr has joined #openstack-manila | 20:17 | |
| bswartz | tbarron: http://stackoverflow.com/questions/1303347/getting-a-map-to-return-a-list-in-python-3-x | 20:18 |
| tbarron | bswartz: back | 20:18 |
| * tbarron cracks a beer in honor of jungleboyj and netapp beerbash | 20:19 | |
| tbarron | bswartz: ah, it's your evil cute lambda | 20:19 |
| bswartz | wasn't trying to be cute honestly | 20:19 |
| bswartz | I didn't see another way to do it | 20:19 |
| jungleboyj | Nice! I am having a beer right now tbarron . | 20:20 |
| tbarron | jungleboyj: "a beer" in the sense of one or more I expect | 20:21 |
| tbarron | I've heard about the lenovo life ... | 20:22 |
| *** markstur has joined #openstack-manila | 20:22 | |
| jungleboyj | tbarron: out with my dad. I am technically on vacation today. | 20:24 |
| *** markstur has quit IRC | 20:28 | |
| bswartz | tbarron: so you're saying tox with no args runs all the same things the gate would? (minus tempest) | 20:32 |
| tbarron | jungleboyj: yeah, heard you are on vacation. Go have fun! | 20:33 |
| tbarron | bswartz: it doesn't run nonvoting cover, etc. | 20:33 |
| tbarron | the defaults for plain 'tox' cmd are at the top of tox.ini | 20:34 |
| tbarron | extra envs are defined in there as well, e.g. to build releasenotes, docs, etc. | 20:34 |
| tbarron | just run them with 'tox -e<env>' | 20:34 |
| tbarron | I rather like the way it sets up venvs and checks dependencies/requirements. But if you just want to run a single test it takes longer than run_tests.sh because it does that every time. | 20:35 |
| tbarron | It keeps your old venvs, but it checks. | 20:36 |
| bswartz | I see | 20:37 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: [WIP] Allow 2 or more export IPs for LVM driver https://review.openstack.org/444479 | 20:37 |
| bswartz | fixed py35 ^ | 20:38 |
| jungleboyj | tbarron: i am having fun. :-) But now I have to go help my parents clean out their basement. | 20:39 |
| tbarron | jay is a good jungleboyj | 20:39 |
| jungleboyj | tbarron: I do my best! | 20:41 |
| jungleboyj | Especially when given lunch and beer first. | 20:41 |
| tbarron | bswartz: yeah I did list(map(....)) everywhere and py35 is happy | 20:42 |
| *** kaisers_ has joined #openstack-manila | 20:44 | |
| bswartz | tbarron: it feels like pyunit should be smart enough to listify map objects itself | 20:47 |
| bswartz | tbarron: we should add an assertListEquals() to test.TestCase I guess | 20:48 |
| *** kaisers_ has quit IRC | 20:49 | |
| *** lpetrut has joined #openstack-manila | 20:53 | |
| *** dustins has joined #openstack-manila | 20:56 | |
| *** eharney has quit IRC | 21:07 | |
| tbarron | bswartz: hmm, we have assertSequenceEqual | 21:11 |
| bswartz | tbarron: where? | 21:11 |
| tbarron | and assertItemsEqual when you don't care about order | 21:12 |
| tbarron | we just pick them up from python unittests I think | 21:12 |
| *** cknight has quit IRC | 21:13 | |
| tbarron | manila/tests/share/drivers/netapp/dataontap/client/test_client_cmode.py | 21:14 |
| tbarron | since it's beer bash time I use that example ^^ | 21:15 |
| *** porrua has quit IRC | 21:15 | |
| bswartz | tbarron: AssertionError: First sequence has no length. Non-sequence?' | 21:15 |
| bswartz | no magic there | 21:16 |
| tbarron | bswartz: dunno, just saw the asserts and thought they *should* work | 21:19 |
| tbarron | oh, they don't work with generators | 21:19 |
| tbarron | is your list already created before the assert? | 21:20 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: [WIP] Allow 2 or more export IPs for LVM driver https://review.openstack.org/444479 | 21:20 |
| bswartz | tbarron: how about this version? | 21:20 |
| tbarron | lookin | 21:20 |
| bswartz | tbarron I passed the map object down to assertSequenceEqual() and it blew up because the map object didn't have a "length" | 21:20 |
| tbarron | yeah, it's like the generator case discussed here: http://stackoverflow.com/questions/7473071/unittests-assertequal-and-iterables-only-check-the-contents (igore the title of the link) Putting list() around it before doing the comparisons fixes that part I guess. | 21:24 |
| bswartz | it's curious that py27's map returns a proper list when an iterator is obviously better | 21:25 |
| bswartz | maybe it's something they realized but couldn't fix in py2 so they changed it in py3 | 21:26 |
| tbarron | Your comparison tests that the lists are the same object, right? assertEqual vs assertSequenceEqual ? | 21:26 |
| bswartz | big language changes are hard to make | 21:26 |
| bswartz | tbarron: I honestly don't know how assertSequenceEqual is different from assertEqual except that it obviously expects the things passed in to have lengths | 21:27 |
| bswartz | if you have 2 proper lists already it's not clear how that's better than plain assertEquals | 21:28 |
| bswartz | what counts as a sequence but not a list? | 21:28 |
| bswartz | an ordered set? | 21:29 |
| bswartz | infinite-length generators? | 21:29 |
| bswartz | how would you test 2 infinite sequences for equality? | 21:30 |
| * bswartz having flashbacks to set theory class | 21:30 | |
| tbarron | the link I cited says assertSequenceEqual can't do generators (infinite sequences effectively) so that's not it | 21:30 |
| tbarron | sequences include immutable tuples as well as mutable lists | 21:31 |
| tbarron | but that still begs the question of why assertEqual isn't sufficient once you have finite sequence | 21:31 |
| * tbarron is no python language wizard | 21:32 | |
| *** markstur has joined #openstack-manila | 21:41 | |
| bswartz | tbarron: iterators are finite but you don't know that until they stop yielding values | 21:43 |
| tbarron | bswartz: :D | 21:43 |
| bswartz | and by that time you've already made a list out of them | 21:43 |
| bswartz | I'm happy with a special list comparison | 21:44 |
| bswartz | it's cleaner than listifying all the maps | 21:44 |
| tbarron | it's certainly cleaner for this patch | 21:44 |
| bswartz | gah! pep8 :-( | 21:45 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila master: [WIP] Allow 2 or more export IPs for LVM driver https://review.openstack.org/444479 | 21:45 |
| tbarron | so we'll also need a patch for devstack so that the lvm job sees the new list config value | 21:47 |
| tbarron | currently devstack/settings sets MANILA_SHARE_EXPORT_IP from $HOST_IP and | 21:48 |
| tbarron | plugin.sh does 'iniset $MANILA_CONF $backend_name lvm_share_export_ip $MANILA_LVM_SHARE_EXPORT_IP' | 21:49 |
| tbarron | one way to do it would be just to | 21:49 |
| tbarron | set MANILA_SHARE_EXPORT_IPV6 from $HOST_IPV6 (which is already getting set w/o doing anything special) | 21:49 |
| tbarron | in settings | 21:50 |
| tbarron | and then combine MANILA_SHARE_EXPORT_IP (or IPV4 if we rename it) | 21:50 |
| tbarron | with MANILA_SHARE_EXPORT_IPV6 in plugin.sh before doing an iniset of lvm_share_export_ips (plural) | 21:51 |
| *** gouthamr has quit IRC | 22:03 | |
| tbarron | bswartz: I'm signing off for now and will be on vacation until next Tuesday. Nice work on the lvm and its helpers! Feel like progress is being made ... | 22:04 |
| *** lpetrut has quit IRC | 22:06 | |
| *** lpetrut has joined #openstack-manila | 22:07 | |
| *** lpetrut has quit IRC | 22:11 | |
| *** dustins has quit IRC | 22:12 | |
| *** eharney has joined #openstack-manila | 22:16 | |
| *** chlong_ has quit IRC | 22:21 | |
| *** catintheroof has quit IRC | 22:23 | |
| *** ganso has quit IRC | 22:25 | |
| *** gouthamr has joined #openstack-manila | 22:29 | |
| *** kaisers_ has joined #openstack-manila | 22:40 | |
| *** kaisers_ has quit IRC | 22:44 | |
| *** porrua has joined #openstack-manila | 23:03 | |
| *** porrua_ has joined #openstack-manila | 23:23 | |
| *** porrua has quit IRC | 23:27 | |
| *** xyang1 has quit IRC | 23:33 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!