| *** emagana has quit IRC | 00:01 | |
| *** ebalduf has joined #openstack-cinder | 00:03 | |
| *** annegentle has quit IRC | 00:04 | |
| *** annegentle has joined #openstack-cinder | 00:09 | |
| openstackgerrit | wangxiyuan proposed openstack/cinder: Fix wrong response with version details https://review.openstack.org/185570 | 00:15 |
|---|---|---|
| *** markvoelker has joined #openstack-cinder | 00:16 | |
| *** winston-d_ has joined #openstack-cinder | 00:17 | |
| *** garthb__ has quit IRC | 00:19 | |
| *** garthb has quit IRC | 00:19 | |
| *** david-lyle has quit IRC | 00:20 | |
| *** annegentle has quit IRC | 00:21 | |
| *** markvoelker has quit IRC | 00:23 | |
| *** winston-d_ has quit IRC | 00:23 | |
| *** markvoelker has joined #openstack-cinder | 00:23 | |
| *** winston-d_ has joined #openstack-cinder | 00:23 | |
| *** annegentle has joined #openstack-cinder | 00:27 | |
| *** Apoorva has quit IRC | 00:27 | |
| *** Apoorva has joined #openstack-cinder | 00:28 | |
| openstackgerrit | xing-yang proposed openstack/cinder: Add flag assume_virtual_capacity_consumed https://review.openstack.org/185764 | 00:29 |
| winston-d | patrickeast, xyang1: I'd echo what Patrick and Sean had said, to me 2 x 10TB and 10 x 2TB are identical in terms of capacity management | 00:29 |
| winston-d | xyang1: it'd help if you can provide an example to demo that they don't. | 00:30 |
| *** Apoorva has quit IRC | 00:31 | |
| openstackgerrit | Patrick East proposed openstack/cinder-specs: Generic image volume cache functionality https://review.openstack.org/182520 | 00:33 |
| xyang1 | winston-d: Not sure what you guys are saying,I just try to accommodate two requirememts here, have you got a chance to look at my patch | 00:34 |
| winston-d | xyang1: "what's the difference between provisioning 200TB of space on 100TB physical whether it is 100 2TB volumes or 1 200 TB volume?" | 00:35 |
| *** annegentle has quit IRC | 00:36 | |
| xyang1 | winston-d: I am not trying to make that difference, it depends on the total capacity, free capacity, your usage and over subscription ratio, I can't say whether you can provision them or not | 00:37 |
| *** daneyon has joined #openstack-cinder | 00:38 | |
| patrickeast | xyang1: i think we are all trying to get the same thing here… the problem i was looking to fix with that bug is that right now we *do* make that distinction | 00:39 |
| patrickeast | xyang1: we would allow 100 2TB volumes but not 1 200TB volumes | 00:39 |
| winston-d | exactly | 00:39 |
| winston-d | by we, that means current cinder CapacityFilter | 00:40 |
| xyang1 | patrickeast: You'll have to lay out the complete formula | 00:40 |
| *** heyun has joined #openstack-cinder | 00:41 | |
| patrickeast | xyang1: so the formula you have already in there for the over subscription works exactly right, but if we go into that first ‘if free < volume_size we return before ever doing any of the virtual capacity calculations | 00:41 |
| *** daneyon has quit IRC | 00:43 | |
| xyang1 | patrickeast: Can you take a look of my latest patch set | 00:44 |
| patrickeast | xyang1: i think the latest patchset you posted is doing the right thing now, it will skip the ‘free < volume_size’ check and make it to the vritual stuff | 00:44 |
| patrickeast | xyang1: yep was just looking at it | 00:44 |
| xyang1 | patrickeast: The existing code was based a review suggestion to be conservative | 00:45 |
| xyang1 | patrickeast: It is hard to accommodate opposite comments | 00:46 |
| xyang1 | patrickeast: I hope the flag can accommodate both | 00:46 |
| *** zhenguo has joined #openstack-cinder | 00:46 | |
| patrickeast | xyang1: yea i had assumed that was probably the case, no worries, i think like most things in cinder everyone has their own way of doing it | 00:46 |
| patrickeast | xyang1: i am curious what the argument is for it though, hopefully i’m not missing something and advocating for a bad change :o | 00:47 |
| xyang1 | patrickeast: That comment came right after the Paris summit | 00:48 |
| winston-d | https://review.openstack.org/#/c/142171/ the change | 00:48 |
| xyang1 | patrickeast: I don't even remember who made that comment. I may be able to dig out that long email if I try hard enough:) | 00:49 |
| xyang1 | patrickeast: After the code was merged though, all the comments I got was similar to yours. That is why I thought about using a flag | 00:50 |
| *** _cjones_ has quit IRC | 00:50 | |
| xyang1 | winston-d: I think that comment was from an email to me | 00:51 |
| winston-d | patrickeast: maybe you can submit a fix _without_ flag, and we have everybody look at two version of fixes and hopefully get some consenus in next cinder weekly meeting? | 00:53 |
| xyang1 | winston-d: Why not just have everyone look at my patch? | 00:54 |
| winston-d | personally, i think there is not difference between 100 2TB and 1 200TB | 00:57 |
| winston-d | so why do we need a flag? | 00:57 |
| patrickeast | hmm so i found where it came in https://review.openstack.org/#/c/142171/19..20/cinder/scheduler/filters/capacity_filter.py but i don’t see comments around patchset 18 or 19 for it | 00:58 |
| xyang1 | winston-d: That was coming from a comment as I said. If everyone thinks it is not right, may be we need to think about it again | 00:59 |
| *** smoriya has joined #openstack-cinder | 00:59 | |
| winston-d | xyang1: i can leave same comment to your fix, i'm fine with that. | 00:59 |
| xyang1 | winston-d: patrickeast I'll try to dig out the email, the person who made the suggestion said that is what they do in practice | 01:00 |
| patrickeast | xyang1: winston-d: we maybe lets leave xing’s patch up for now and see what others think | 01:00 |
| patrickeast | xyang1: at the meeting* | 01:00 |
| patrickeast | xyang1: oh interesting, so maybe some backends can’t support it | 01:00 |
| xyang1 | winston-d: patrickeast sounds good. | 01:01 |
| xyang1 | patrickeast: Not sure, it could just be operator is cautious | 01:01 |
| winston-d | just in case I miss the meeting this week (it always happen 1st week when i start work on US timezone), I will vote for not adding new flag. | 01:02 |
| xyang1 | winston-d: Ok | 01:02 |
| xyang1 | winston-d: That is fine with our array | 01:02 |
| winston-d | if it's backend specific limit, it shouldn't be solved by a global flag. | 01:02 |
| *** asselin_ has joined #openstack-cinder | 01:04 | |
| *** xiaohui has joined #openstack-cinder | 01:05 | |
| xyang1 | winston-d: Can you look at the rest of the patch? Line 291 of host_manager.py | 01:05 |
| xyang1 | winston-d: We can add the flag to backend specific section | 01:05 |
| *** tobe has joined #openstack-cinder | 01:09 | |
| *** dims has quit IRC | 01:12 | |
| *** patrickeast has quit IRC | 01:14 | |
| *** IlyaG has quit IRC | 01:16 | |
| *** jamielennox|away is now known as jamielennox | 01:23 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 01:24 |
| *** annegentle has joined #openstack-cinder | 01:27 | |
| *** annegentle has quit IRC | 01:28 | |
| *** Apoorva has joined #openstack-cinder | 01:29 | |
| *** Lee1092 has joined #openstack-cinder | 01:30 | |
| *** setmason has quit IRC | 01:35 | |
| *** david-lyle has joined #openstack-cinder | 01:35 | |
| *** annegentle has joined #openstack-cinder | 01:49 | |
| *** rushil has joined #openstack-cinder | 02:07 | |
| *** Apoorva has quit IRC | 02:09 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix response when querying host detail by host name https://review.openstack.org/162601 | 02:09 |
| *** rushil has quit IRC | 02:10 | |
| nikeshm | asselin: hi wanted to extend my setup to multiple providers,nodepool image-build is working good but image-upload is failing http://paste.openstack.org/show/238941/ | 02:15 |
| *** tbarron1 has quit IRC | 02:16 | |
| nikeshm | asselin: this my nodepool yaml file http://paste.openstack.org/show/238951/ | 02:17 |
| nikeshm | patrickeast: hi | 02:19 |
| *** akerr_ has quit IRC | 02:20 | |
| *** daneyon has joined #openstack-cinder | 02:27 | |
| asselin_ | nikeshm, I haven't seen that error before | 02:28 |
| asselin_ | nikeshm, look like something related to 'shade'. This is new | 02:28 |
| asselin_ | nikeshm, looks like it merged recently: https://review.openstack.org/#/c/168633/ | 02:29 |
| *** daneyon has quit IRC | 02:31 | |
| asselin_ | nikeshm, just asked in -infra. Issue is fixed in a recent release of shade | 02:32 |
| *** david-lyle has quit IRC | 02:33 | |
| *** setmason has joined #openstack-cinder | 02:33 | |
| asselin_ | nikeshm, I don't have my env, but re-running ~./install_master.sh should pull in the latest | 02:34 |
| asselin_ | nikeshm, if not, go to /opt/shade | 02:35 |
| asselin_ | sudo git pull | 02:35 |
| asselin_ | sudo pip install . | 02:35 |
| asselin_ | and try again | 02:35 |
| *** annegentle has quit IRC | 02:38 | |
| *** Yogi1 has joined #openstack-cinder | 02:39 | |
| *** annegentle has joined #openstack-cinder | 02:40 | |
| *** asselin_ has quit IRC | 02:54 | |
| nikeshm | asselin: there is no /opt/shade | 03:02 |
| nikeshm | also tried to rerun install master | 03:02 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 03:04 |
| *** tobe has quit IRC | 03:08 | |
| *** david-lyle has joined #openstack-cinder | 03:11 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 03:17 |
| *** dmz has joined #openstack-cinder | 03:18 | |
| dmz | howdy y'all, i thought i had cinder working but was just trying to attach a volume & got the following in the nova-api log: ERROR nova.api.openstack [req-66ecda54-cddc-48f2-a58c-4ff49a7b056a None] Caught error: AmbiguousEndpoints, ; looks like it's uing an authv1 token but not sure what this problem could be | 03:19 |
| *** kaisers has quit IRC | 03:20 | |
| *** tobe has joined #openstack-cinder | 03:26 | |
| *** Yogi1 has quit IRC | 03:31 | |
| openstackgerrit | wangchy proposed openstack/cinder: Stop registering opt fatal_exception_format_errors in cinder.exp https://review.openstack.org/185314 | 03:32 |
| *** annegentle has quit IRC | 03:32 | |
| *** changbl has joined #openstack-cinder | 03:32 | |
| *** annegentle has joined #openstack-cinder | 03:37 | |
| *** dalgaaf has quit IRC | 03:38 | |
| *** Apoorva has joined #openstack-cinder | 03:42 | |
| openstackgerrit | Alex Meade proposed openstack/cinder: NetApp E-Series driver: Remove caching logic https://review.openstack.org/185832 | 03:42 |
| openstackgerrit | Alex Meade proposed openstack/cinder: NetApp E-Series: Refactor class structure for FC https://review.openstack.org/185833 | 03:42 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 03:44 |
| *** diga has joined #openstack-cinder | 03:51 | |
| *** emagana has joined #openstack-cinder | 03:51 | |
| *** tobe has quit IRC | 03:58 | |
| openstackgerrit | xing-yang proposed openstack/os-brick: Add connector driver for the ScaleIO cinder driver https://review.openstack.org/185835 | 03:58 |
| *** Apoorva has quit IRC | 03:59 | |
| *** krtaylor has joined #openstack-cinder | 04:04 | |
| *** daneyon has joined #openstack-cinder | 04:15 | |
| *** links has joined #openstack-cinder | 04:15 | |
| *** daneyon has quit IRC | 04:20 | |
| *** xyang1 has quit IRC | 04:20 | |
| *** annegentle has quit IRC | 04:26 | |
| *** markvoelker has quit IRC | 04:31 | |
| *** ebalduf has quit IRC | 04:33 | |
| *** tobe has joined #openstack-cinder | 04:34 | |
| *** avishay__ has joined #openstack-cinder | 04:35 | |
| *** avishay__ is now known as avishay | 04:36 | |
| *** sks has joined #openstack-cinder | 04:39 | |
| *** bkopilov is now known as bkopilov_wfh | 04:52 | |
| *** mdbooth has quit IRC | 04:55 | |
| *** pradipta has joined #openstack-cinder | 04:55 | |
| *** harlowja_at_home has joined #openstack-cinder | 04:57 | |
| *** mdbooth has joined #openstack-cinder | 04:59 | |
| dmz | :( | 05:02 |
| *** tobe has quit IRC | 05:04 | |
| *** nkrinner has joined #openstack-cinder | 05:05 | |
| openstackgerrit | wangchy proposed openstack/cinder: Stop registering opt fatal_exception_format_errors in cinder.exp https://review.openstack.org/185314 | 05:17 |
| *** leopoldj has joined #openstack-cinder | 05:18 | |
| openstackgerrit | Ph. Marek proposed openstack/cinder: Re-add DRBD driver. https://review.openstack.org/178573 | 05:23 |
| *** belmoreira has joined #openstack-cinder | 05:23 | |
| *** harlowja_at_home has quit IRC | 05:24 | |
| *** avishay has quit IRC | 05:25 | |
| *** avishay has joined #openstack-cinder | 05:25 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 05:27 |
| *** deepakcs has joined #openstack-cinder | 05:28 | |
| *** cloudm2 has quit IRC | 05:32 | |
| *** setmason has quit IRC | 05:34 | |
| *** setmason has joined #openstack-cinder | 05:35 | |
| *** IanGovett has joined #openstack-cinder | 05:44 | |
| *** smoriya has quit IRC | 05:46 | |
| *** tobe has joined #openstack-cinder | 05:50 | |
| *** IanGovett has quit IRC | 06:00 | |
| *** emagana has quit IRC | 06:00 | |
| *** emagana has joined #openstack-cinder | 06:01 | |
| *** kaisers has joined #openstack-cinder | 06:01 | |
| *** e0ne has joined #openstack-cinder | 06:01 | |
| *** kaisers has quit IRC | 06:02 | |
| *** daneyon has joined #openstack-cinder | 06:04 | |
| *** emagana has quit IRC | 06:05 | |
| *** kaisers has joined #openstack-cinder | 06:06 | |
| *** BharatK has joined #openstack-cinder | 06:08 | |
| *** daneyon has quit IRC | 06:09 | |
| *** e0ne has quit IRC | 06:11 | |
| openstackgerrit | XinXiaohui proposed openstack/cinder-specs: capacity-headroom https://review.openstack.org/170380 | 06:13 |
| openstackgerrit | wanghao proposed openstack/cinder: Notify the transfer volume action in cinder https://review.openstack.org/185531 | 06:16 |
| *** tobe has quit IRC | 06:20 | |
| *** ankit_ag has joined #openstack-cinder | 06:22 | |
| openstackgerrit | Julien Danjou proposed openstack/cinder: Stop using deprecated timeutils.isotime() https://review.openstack.org/184330 | 06:28 |
| *** kaisers has quit IRC | 06:30 | |
| openstackgerrit | Ph. Marek proposed openstack/cinder: Re-add DRBD driver. https://review.openstack.org/178573 | 06:31 |
| *** tobe has joined #openstack-cinder | 06:38 | |
| *** avishay_ has joined #openstack-cinder | 06:51 | |
| *** avishay has quit IRC | 06:52 | |
| *** lpetrut has joined #openstack-cinder | 06:55 | |
| *** davechen1 has joined #openstack-cinder | 06:57 | |
| *** agarciam has joined #openstack-cinder | 07:00 | |
| *** setmason has quit IRC | 07:06 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 07:07 |
| *** setmason has joined #openstack-cinder | 07:12 | |
| *** jordanP has joined #openstack-cinder | 07:14 | |
| *** IanGovett has joined #openstack-cinder | 07:17 | |
| *** markvoelker has joined #openstack-cinder | 07:17 | |
| *** alecv has joined #openstack-cinder | 07:19 | |
| *** bkopilov_wfh has quit IRC | 07:20 | |
| *** dulek has joined #openstack-cinder | 07:20 | |
| *** ronis has joined #openstack-cinder | 07:22 | |
| *** anshul has joined #openstack-cinder | 07:25 | |
| *** emagana has joined #openstack-cinder | 07:25 | |
| *** bkopilov has joined #openstack-cinder | 07:25 | |
| *** annegentle has joined #openstack-cinder | 07:27 | |
| openstackgerrit | Abhijeet Malawade proposed openstack/cinder: Remove unused context parameter https://review.openstack.org/185861 | 07:29 |
| *** emagana has quit IRC | 07:30 | |
| *** annegentle has quit IRC | 07:32 | |
| *** markus_z has joined #openstack-cinder | 07:32 | |
| *** avishay_ has quit IRC | 07:32 | |
| *** avishay_ has joined #openstack-cinder | 07:33 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 07:35 |
| *** sgotliv has joined #openstack-cinder | 07:36 | |
| *** tshefi has joined #openstack-cinder | 07:36 | |
| *** tobe has quit IRC | 07:37 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 07:37 |
| *** dims has joined #openstack-cinder | 07:38 | |
| *** tobe has joined #openstack-cinder | 07:38 | |
| *** setmason has quit IRC | 07:41 | |
| *** dims has quit IRC | 07:44 | |
| *** chlong has quit IRC | 07:45 | |
| *** dalgaaf has joined #openstack-cinder | 07:48 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 07:51 |
| *** jistr has joined #openstack-cinder | 07:52 | |
| *** daneyon has joined #openstack-cinder | 07:53 | |
| openstackgerrit | Michal Dulko proposed openstack/cinder: Backup object https://review.openstack.org/157085 | 07:54 |
| openstackgerrit | Michal Dulko proposed openstack/cinder: Service object https://review.openstack.org/160417 | 07:55 |
| *** daneyon has quit IRC | 07:58 | |
| *** emagana has joined #openstack-cinder | 08:04 | |
| *** emagana has quit IRC | 08:09 | |
| *** e0ne has joined #openstack-cinder | 08:15 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 08:15 |
| *** e0ne has quit IRC | 08:19 | |
| openstackgerrit | Liu Xinguo proposed openstack/cinder: Fix the ranges in RestURL with Huawei drivers https://review.openstack.org/173157 | 08:23 |
| *** ndipanov has joined #openstack-cinder | 08:36 | |
| *** heyun has quit IRC | 08:36 | |
| *** markus_z has quit IRC | 08:38 | |
| *** tobe has quit IRC | 08:42 | |
| *** tobe has joined #openstack-cinder | 08:43 | |
| *** lpetrut has quit IRC | 08:46 | |
| *** markus_z has joined #openstack-cinder | 08:51 | |
| openstackgerrit | Abhijeet Malawade proposed openstack/cinder: Remove unused context parameter https://review.openstack.org/185861 | 08:51 |
| openstackgerrit | Xi Yang proposed openstack/cinder: Provide snap copy feature in EMC VNX Cinder driver https://review.openstack.org/184733 | 08:54 |
| *** gardenshed has joined #openstack-cinder | 08:54 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Create iSCSI lio portals with right IPs and port https://review.openstack.org/161209 | 08:55 |
| *** ebalduf has joined #openstack-cinder | 08:57 | |
| *** emagana has joined #openstack-cinder | 08:59 | |
| *** avishay_ has quit IRC | 09:00 | |
| *** lifeless has quit IRC | 09:00 | |
| *** lifeless_ has joined #openstack-cinder | 09:00 | |
| *** gardenshed has quit IRC | 09:01 | |
| *** daneyon has joined #openstack-cinder | 09:01 | |
| *** ebalduf has quit IRC | 09:02 | |
| *** emagana has quit IRC | 09:03 | |
| *** avishay_ has joined #openstack-cinder | 09:03 | |
| *** daneyon has quit IRC | 09:06 | |
| *** belmoreira has quit IRC | 09:15 | |
| *** ekarlso has quit IRC | 09:16 | |
| *** rushiagr_away is now known as rushiagr | 09:18 | |
| openstackgerrit | Liu Xinguo proposed openstack/cinder: Fix the ranges in RestURL with Huawei drivers https://review.openstack.org/173157 | 09:22 |
| wanghao | DuncanT: ping | 09:22 |
| wanghao | DuncanT: Would you have a look to this bp: https://review.openstack.org/#/c/180400/. Is there something else that we need to fix? | 09:22 |
| *** ekarlso has joined #openstack-cinder | 09:22 | |
| *** e0ne has joined #openstack-cinder | 09:23 | |
| *** e0ne is now known as e0ne_ | 09:23 | |
| DuncanT | wanghao: The two comments on https://review.openstack.org/#/c/180400/17/cinder/volume/utils.py could do with looking at... Sorry it is taking so long to get through, it happens some times. Both are reasonable comments | 09:24 |
| wanghao | DuncanT: Well, may be we need one more patch for this. I will do it soon. Thanks:) | 09:26 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 09:27 |
| *** sgotliv has quit IRC | 09:27 | |
| DuncanT | wanghao: Ping me when it's up and I should be able to get the first +2 on there quickly | 09:28 |
| wanghao | DuncanT: Sure | 09:29 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 09:30 |
| *** e0ne_ is now known as e0ne | 09:31 | |
| *** Liu has joined #openstack-cinder | 09:32 | |
| *** lpetrut has joined #openstack-cinder | 09:33 | |
| *** diga has quit IRC | 09:33 | |
| *** theanalyst has quit IRC | 09:34 | |
| *** dosaboy_ has quit IRC | 09:36 | |
| *** dosaboy has joined #openstack-cinder | 09:36 | |
| *** theanalyst has joined #openstack-cinder | 09:37 | |
| *** kaisers has joined #openstack-cinder | 09:39 | |
| openstackgerrit | wanghao proposed openstack/cinder: Notification with volume and snapshot metadata https://review.openstack.org/180400 | 09:42 |
| openstackgerrit | rakesh mishra proposed openstack/cinder: cinder-quota-define-per-volume https://review.openstack.org/185906 | 09:46 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 09:48 |
| *** nikeshm has quit IRC | 09:48 | |
| *** davechen1 has left #openstack-cinder | 09:48 | |
| *** gardenshed has joined #openstack-cinder | 09:51 | |
| *** emagana has joined #openstack-cinder | 09:53 | |
| *** Arkady_Kanevsky has quit IRC | 09:56 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 09:57 |
| *** emagana has quit IRC | 09:57 | |
| *** theanalyst has quit IRC | 10:02 | |
| *** theanalyst has joined #openstack-cinder | 10:04 | |
| *** rushiagr has quit IRC | 10:08 | |
| *** theanalyst has quit IRC | 10:08 | |
| *** dims has joined #openstack-cinder | 10:09 | |
| wanghao | DuncanT: ping | 10:12 |
| wanghao | Duncant: The patch has been commited. You can have a look after jenkins passed. Thanks. | 10:12 |
| *** theanalyst has joined #openstack-cinder | 10:12 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 10:15 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Add column "hidden" to volumes table https://review.openstack.org/185857 | 10:26 |
| openstackgerrit | Szymon Wróblewski proposed openstack/cinder: Tooz locks https://review.openstack.org/183537 | 10:35 |
| openstackgerrit | Michal Dulko proposed openstack/cinder: Service object https://review.openstack.org/160417 | 10:36 |
| dulek | DuncanT: Hi, do we want to add A/A cinder-volume to today's meeting agenda? | 10:38 |
| *** nlevinki has joined #openstack-cinder | 10:42 | |
| *** emagana has joined #openstack-cinder | 10:47 | |
| *** emagana has quit IRC | 10:48 | |
| *** emagana has joined #openstack-cinder | 10:48 | |
| *** pradipta has quit IRC | 10:49 | |
| *** daneyon has joined #openstack-cinder | 10:50 | |
| openstackgerrit | Michal Dulko proposed openstack/cinder: POC: Tooz locks demo https://review.openstack.org/185646 | 10:51 |
| *** emagana has quit IRC | 10:53 | |
| *** daneyon has quit IRC | 10:55 | |
| *** dulek_ has joined #openstack-cinder | 10:57 | |
| *** dulek has quit IRC | 10:59 | |
| *** rushiagr_away has joined #openstack-cinder | 11:00 | |
| openstackgerrit | Pavel Boldin proposed openstack/cinder: Add iscsi_target_flags configuration option https://review.openstack.org/182871 | 11:00 |
| *** ebalduf has joined #openstack-cinder | 11:09 | |
| *** ebalduf has quit IRC | 11:13 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Volume get and list should be able get detailed views https://review.openstack.org/156292 | 11:15 |
| *** sgotliv has joined #openstack-cinder | 11:15 | |
| *** e0ne is now known as e0ne_ | 11:26 | |
| *** openstack has joined #openstack-cinder | 11:38 | |
| *** emagana has joined #openstack-cinder | 11:43 | |
| *** aix has joined #openstack-cinder | 11:47 | |
| *** emagana has quit IRC | 11:48 | |
| *** tobe has joined #openstack-cinder | 11:52 | |
| *** deepakcs has quit IRC | 11:56 | |
| *** tobe has quit IRC | 11:56 | |
| *** tobe has joined #openstack-cinder | 11:57 | |
| *** dulek___ has quit IRC | 11:58 | |
| *** dulek___ has joined #openstack-cinder | 11:58 | |
| *** BharatK has quit IRC | 12:01 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Create iSCSI lio portals with right IPs and port https://review.openstack.org/161209 | 12:01 |
| *** markvoelker has quit IRC | 12:04 | |
| *** markvoelker has joined #openstack-cinder | 12:04 | |
| *** dulek_ has joined #openstack-cinder | 12:06 | |
| *** bkopilov is now known as bkopilov_wfh | 12:08 | |
| *** akerr has joined #openstack-cinder | 12:09 | |
| *** e0ne is now known as e0ne_ | 12:09 | |
| *** dulek___ has quit IRC | 12:09 | |
| *** e0ne_ is now known as e0ne | 12:14 | |
| *** Yogi1 has joined #openstack-cinder | 12:14 | |
| *** BharatK has joined #openstack-cinder | 12:15 | |
| *** avishay_ is now known as avishay | 12:15 | |
| *** bswartz has quit IRC | 12:17 | |
| *** BharatK has quit IRC | 12:19 | |
| *** ebalduf has joined #openstack-cinder | 12:29 | |
| *** dims has quit IRC | 12:31 | |
| *** dims has joined #openstack-cinder | 12:32 | |
| *** dulek_ has quit IRC | 12:32 | |
| *** ebalduf has quit IRC | 12:33 | |
| *** emagana has joined #openstack-cinder | 12:37 | |
| *** daneyon has joined #openstack-cinder | 12:39 | |
| *** akshai has joined #openstack-cinder | 12:40 | |
| *** eharney has quit IRC | 12:40 | |
| *** emagana has quit IRC | 12:41 | |
| *** annegentle has joined #openstack-cinder | 12:41 | |
| *** daneyon has quit IRC | 12:44 | |
| *** kaisers has quit IRC | 12:45 | |
| *** Yogi1 has quit IRC | 12:45 | |
| *** julim has joined #openstack-cinder | 12:48 | |
| *** marcusvrn has joined #openstack-cinder | 12:52 | |
| *** ekarlso has quit IRC | 12:53 | |
| *** ekarlso has joined #openstack-cinder | 12:53 | |
| *** Yogi1 has joined #openstack-cinder | 12:54 | |
| *** xyang1 has joined #openstack-cinder | 12:56 | |
| *** sks has quit IRC | 12:56 | |
| *** dulek has joined #openstack-cinder | 12:58 | |
| *** Yogi1 has quit IRC | 12:58 | |
| *** nlevinki_ has joined #openstack-cinder | 13:03 | |
| *** nlevinki has quit IRC | 13:05 | |
| *** annegentle has quit IRC | 13:06 | |
| *** simondodsley has joined #openstack-cinder | 13:09 | |
| *** lpetrut has quit IRC | 13:10 | |
| *** sks has joined #openstack-cinder | 13:10 | |
| *** mriedem_away is now known as mriedem | 13:17 | |
| *** juzuluag has joined #openstack-cinder | 13:17 | |
| *** rasoto_ has joined #openstack-cinder | 13:20 | |
| *** lpetrut has joined #openstack-cinder | 13:23 | |
| *** alonmarx_ has joined #openstack-cinder | 13:23 | |
| *** Yogi1 has joined #openstack-cinder | 13:25 | |
| *** dustins has joined #openstack-cinder | 13:26 | |
| *** alonmarx has quit IRC | 13:26 | |
| *** annegentle has joined #openstack-cinder | 13:28 | |
| *** eharney has joined #openstack-cinder | 13:30 | |
| *** abhiram_moturi has joined #openstack-cinder | 13:30 | |
| *** annegentle has quit IRC | 13:30 | |
| *** emagana has joined #openstack-cinder | 13:31 | |
| *** annegentle has joined #openstack-cinder | 13:33 | |
| *** timcl has joined #openstack-cinder | 13:34 | |
| *** emagana has quit IRC | 13:35 | |
| *** abhiram_moturi has quit IRC | 13:38 | |
| *** abhiram_moturi has joined #openstack-cinder | 13:39 | |
| *** jungleboyj has quit IRC | 13:39 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Create iSCSI lio portals with right IPs and port https://review.openstack.org/161209 | 13:40 |
| *** links has quit IRC | 13:41 | |
| *** cloudm2 has joined #openstack-cinder | 13:42 | |
| *** abhiram_moturi has quit IRC | 13:43 | |
| *** abhiram_moturi has joined #openstack-cinder | 13:43 | |
| *** abhiram_moturi has quit IRC | 13:48 | |
| *** abhiram_moturi has joined #openstack-cinder | 13:48 | |
| *** abhiram_moturi has quit IRC | 13:50 | |
| *** abhiram_moturi has joined #openstack-cinder | 13:50 | |
| *** annegentle has quit IRC | 13:52 | |
| *** sgotliv has quit IRC | 13:52 | |
| *** annegentle has joined #openstack-cinder | 13:53 | |
| *** rushil has joined #openstack-cinder | 13:55 | |
| *** dustins_ has joined #openstack-cinder | 13:56 | |
| *** cbits has joined #openstack-cinder | 13:56 | |
| *** cbits has left #openstack-cinder | 13:57 | |
| *** sks has quit IRC | 13:58 | |
| *** Yogi11 has joined #openstack-cinder | 13:58 | |
| *** pradipta has joined #openstack-cinder | 13:59 | |
| *** dustins has quit IRC | 14:00 | |
| *** rushil_ has joined #openstack-cinder | 14:00 | |
| *** daneyon has joined #openstack-cinder | 14:00 | |
| *** rushil has quit IRC | 14:00 | |
| *** timcl1 has joined #openstack-cinder | 14:02 | |
| *** Yogi1 has quit IRC | 14:02 | |
| *** timcl has quit IRC | 14:02 | |
| *** sgotliv has joined #openstack-cinder | 14:03 | |
| Liu | hi | 14:04 |
| Liu | Our FC CI always failed with some test cases, like test_minimum_basic_scenario and test_volume_boot_pattern | 14:04 |
| Liu | but iSCSI CI is ok for these test cases | 14:05 |
| *** chlong has joined #openstack-cinder | 14:05 | |
| Liu | who can give me some point | 14:05 |
| Liu | log link: http://182.138.104.27:8088/huawei-18000-fc-dsvm-tempest-full/423/console.html | 14:05 |
| Liu | Thanks very much | 14:06 |
| *** merooney has joined #openstack-cinder | 14:06 | |
| *** dalgaaf has quit IRC | 14:08 | |
| *** dustins_ is now known as dustins | 14:10 | |
| *** dustins is now known as dustins_away | 14:10 | |
| *** bswartz has joined #openstack-cinder | 14:10 | |
| *** Liu is now known as liuxg | 14:11 | |
| *** merooney has quit IRC | 14:13 | |
| *** merooney has joined #openstack-cinder | 14:13 | |
| *** merooney is now known as sdafasdfdsafasdf | 14:14 | |
| *** sdafasdfdsafasdf is now known as merooney | 14:14 | |
| *** tobe has quit IRC | 14:15 | |
| *** annegentle has quit IRC | 14:17 | |
| *** leopoldj has quit IRC | 14:17 | |
| *** merooney has left #openstack-cinder | 14:18 | |
| *** merooney has joined #openstack-cinder | 14:19 | |
| *** breitz has quit IRC | 14:19 | |
| *** breitz has joined #openstack-cinder | 14:20 | |
| *** abhiram_moturi has quit IRC | 14:24 | |
| *** abhiram_moturi has joined #openstack-cinder | 14:25 | |
| *** nlevinki_ has quit IRC | 14:27 | |
| *** ronis has quit IRC | 14:28 | |
| *** Adriano_ has joined #openstack-cinder | 14:28 | |
| *** jungleboyj has joined #openstack-cinder | 14:30 | |
| *** bnemec has quit IRC | 14:30 | |
| *** thangp has joined #openstack-cinder | 14:30 | |
| *** abhiram_moturi has quit IRC | 14:31 | |
| *** abhiram_moturi has joined #openstack-cinder | 14:31 | |
| *** avishay has quit IRC | 14:31 | |
| *** annegentle has joined #openstack-cinder | 14:32 | |
| *** esker has joined #openstack-cinder | 14:33 | |
| *** mtanino has joined #openstack-cinder | 14:33 | |
| *** avishay has joined #openstack-cinder | 14:34 | |
| *** bnemec has joined #openstack-cinder | 14:35 | |
| *** dims has quit IRC | 14:35 | |
| *** daneyon_ has joined #openstack-cinder | 14:37 | |
| *** abhiram_moturi has quit IRC | 14:39 | |
| *** abhiram_moturi has joined #openstack-cinder | 14:40 | |
| *** daneyon has quit IRC | 14:40 | |
| *** sgotliv has quit IRC | 14:40 | |
| *** emagana has joined #openstack-cinder | 14:40 | |
| *** dulek has quit IRC | 14:41 | |
| *** eharney has quit IRC | 14:42 | |
| *** timcl1 has quit IRC | 14:42 | |
| *** jms has joined #openstack-cinder | 14:44 | |
| *** sgotliv has joined #openstack-cinder | 14:46 | |
| *** timcl has joined #openstack-cinder | 14:47 | |
| *** abhiram_moturi has quit IRC | 14:49 | |
| *** abhiram_moturi has joined #openstack-cinder | 14:49 | |
| *** hemnafk is now known as hemna | 14:50 | |
| *** merooney has quit IRC | 14:50 | |
| *** jungleboyj has quit IRC | 14:50 | |
| *** patrickeast has joined #openstack-cinder | 14:51 | |
| *** anshul has quit IRC | 14:51 | |
| *** jungleboyj has joined #openstack-cinder | 14:53 | |
| openstackgerrit | Thang Pham proposed openstack/cinder: Complete switch to snapshot objects https://review.openstack.org/163910 | 14:53 |
| *** dims_ has joined #openstack-cinder | 14:55 | |
| *** abhiram_moturi has quit IRC | 14:58 | |
| openstackgerrit | Patrick East proposed openstack/cinder-specs: Generic image volume cache functionality https://review.openstack.org/182520 | 14:58 |
| *** rmesta has joined #openstack-cinder | 14:58 | |
| *** abhiram_moturi has joined #openstack-cinder | 14:58 | |
| *** abhiram_moturi has quit IRC | 15:00 | |
| *** kvidvans has joined #openstack-cinder | 15:01 | |
| *** abhiram_moturi has joined #openstack-cinder | 15:01 | |
| *** timcl1 has joined #openstack-cinder | 15:01 | |
| *** timcl has quit IRC | 15:03 | |
| *** tsekiyama has joined #openstack-cinder | 15:04 | |
| *** garthb has joined #openstack-cinder | 15:04 | |
| *** garthb__ has joined #openstack-cinder | 15:04 | |
| *** asselin_ has joined #openstack-cinder | 15:04 | |
| *** merooney has joined #openstack-cinder | 15:05 | |
| *** abhiram_moturi has quit IRC | 15:09 | |
| tsekiyama | thingee: If you have a chance, could you approve https://blueprints.launchpad.net/cinder/+spec/hitachi-remove-resource-lock ? | 15:09 |
| tsekiyama | thingee: This is hitachi driver update to remove unnecessary locks. Patches is proposed here: https://review.openstack.org/#/c/185003 | 15:09 |
| *** abhiram_moturi has joined #openstack-cinder | 15:09 | |
| *** bswartz has quit IRC | 15:10 | |
| *** tshefi has quit IRC | 15:11 | |
| *** anshul has joined #openstack-cinder | 15:12 | |
| *** bswartz has joined #openstack-cinder | 15:12 | |
| *** e0ne is now known as e0ne_ | 15:13 | |
| *** tobe has joined #openstack-cinder | 15:15 | |
| *** tswanson_ has quit IRC | 15:19 | |
| *** Swanson has joined #openstack-cinder | 15:19 | |
| *** dulek_home has joined #openstack-cinder | 15:20 | |
| *** tobe has quit IRC | 15:20 | |
| *** merooney has quit IRC | 15:20 | |
| openstackgerrit | Tina Tang proposed openstack/cinder: Create consistgroup from cgsnapshot support in VNX driver https://review.openstack.org/163706 | 15:21 |
| Swanson | Any cores available to look at this one? https://review.openstack.org/#/c/177994/ I've one +2. Just need another. I changed my API object so I'd like this one in. | 15:21 |
| *** e0ne_ is now known as e0ne | 15:22 | |
| *** abhiram_moturi has quit IRC | 15:26 | |
| *** abhiram_moturi has joined #openstack-cinder | 15:27 | |
| *** tsekiyama1 has joined #openstack-cinder | 15:27 | |
| *** tsekiyama has quit IRC | 15:28 | |
| jms | So, mostly have the the 2nd cinder-volume service on a second host staying "up". When trying to create a volume, it fails with "no host". The logs show that the cinder-scheduler on the 1st host (controller), is only looking at the storage defined in that nodes cinder.conf, though the second node shows up in 'service-list' [so, only looking at controller@lvm]. Any pointers on how to get cinder to recognize the second host, and look for the backend that' | 15:30 |
| *** setmason has joined #openstack-cinder | 15:31 | |
| dulek_home | jms: Do you have synchronized times on both hosts? | 15:31 |
| *** eharney has joined #openstack-cinder | 15:32 | |
| jms | dulek_home: Yes. | 15:32 |
| *** leeantho has joined #openstack-cinder | 15:33 | |
| dulek_home | jms: Can you post the logs on paste.openstack.org? | 15:33 |
| *** BharatK has joined #openstack-cinder | 15:34 | |
| jms | Hah it's still there... http://paste.openstack.org/show/238652/ | 15:37 |
| *** vokt has joined #openstack-cinder | 15:37 | |
| *** ronis has joined #openstack-cinder | 15:38 | |
| *** harlowja_at_home has joined #openstack-cinder | 15:40 | |
| *** merooney has joined #openstack-cinder | 15:40 | |
| *** tsekiyam has joined #openstack-cinder | 15:42 | |
| dulek_home | jms: Maybe c-vol on the node has problems with connecting to the AMQP and haven't sent capabilities? | 15:42 |
| dulek_home | jms: If it had the access to the DB then it would be able to bump up DB heartbeat and appear up in cinder service-list. | 15:43 |
| dmz | howdy y'all, i thought i had cinder working but was just trying to attach a volume & got the following in the nova-api log: ERROR nova.api.openstack [req-66ecda54-cddc-48f2-a58c-4ff49a7b056a None] Caught error: AmbiguousEndpoints, ; this started after I added my 2nd region cinder deplyment; and yes i'm using OS_REGION_NAME env variable & command line options on both the nova client and cinder client attach commands; i can create / delete disks just | 15:43 |
| dmz | not attach | 15:43 |
| *** tsekiyama1 has left #openstack-cinder | 15:43 | |
| *** annegentle has quit IRC | 15:47 | |
| *** jwcroppe has quit IRC | 15:48 | |
| *** jwcroppe has joined #openstack-cinder | 15:48 | |
| jms | dulek_home: It does show up in: cinder service-list -- output. | 15:51 |
| *** lpetrut has quit IRC | 15:52 | |
| *** lpetrut has joined #openstack-cinder | 15:52 | |
| *** anshul has quit IRC | 15:53 | |
| *** lifeless_ is now known as lifeless | 15:53 | |
| *** barra204 has quit IRC | 15:54 | |
| jgriffith | do we have folks for a meeting today? | 15:54 |
| jgriffith | thingee: has asked that I stand in for him | 15:54 |
| dulek_home | jgriffith: Agenda is huge, so probably yes | 15:54 |
| jgriffith | dulek_home: :) | 15:54 |
| jgriffith | dulek_home: it keeps growing evrey time I open it | 15:55 |
| dulek_home | jms: Yeah, it would in case of having DB access but no AMQP accesss. | 15:55 |
| *** dannywilson has joined #openstack-cinder | 15:56 | |
| DuncanT | jms: Symptoms also match driver init failure... can you post up the cinder-volume log of the service that isn't getting used? | 15:57 |
| *** jwcroppe has quit IRC | 15:59 | |
| jgriffith | Ok. let's go have a meeting.. | 15:59 |
| *** deepakcs has joined #openstack-cinder | 16:00 | |
| *** emagana has quit IRC | 16:00 | |
| *** julim has quit IRC | 16:00 | |
| *** agarciam has joined #openstack-cinder | 16:01 | |
| *** agarciam1 has quit IRC | 16:01 | |
| *** emagana has joined #openstack-cinder | 16:02 | |
| *** abhiram_moturi has quit IRC | 16:02 | |
| *** xyang has joined #openstack-cinder | 16:03 | |
| *** rajinir has joined #openstack-cinder | 16:03 | |
| *** julim has joined #openstack-cinder | 16:03 | |
| *** dustins_away is now known as dustins | 16:03 | |
| *** adurbin_ has joined #openstack-cinder | 16:04 | |
| *** merooney has quit IRC | 16:06 | |
| *** jistr has quit IRC | 16:07 | |
| jms | Hrmm...think I found a pointer in /var/log/messages (Yes, another log file to remember to check!) ... | 16:09 |
| jms | May 27 11:04:26 smh0005 cinder-scheduler: AttributeError: 'RequestContext' object has no attribute 'timestamp' | 16:09 |
| jms | May 27 11:04:26 smh0005 cinder-scheduler: Logged from file dispatcher.py, line 151 | 16:09 |
| jms | smh0005 would be the host that has LVM and is being checked. | 16:09 |
| jms | i.e. controller | 16:10 |
| *** asselin__ has joined #openstack-cinder | 16:11 | |
| jms | DuncanT: For volume drive it's getting to: 2015-05-27 09:53:57.607 20835 INFO cinder.volume.manager [req-56fff6f6-f11f-4591-8fb7-414da435b604 - - - - -] Driver initialization completed successfully. | 16:11 |
| jms | Need more of volume.log from that node? | 16:11 |
| *** asselin_ has quit IRC | 16:12 | |
| *** harlowja_at_home has quit IRC | 16:12 | |
| *** rushil_ has quit IRC | 16:14 | |
| *** nikeshm has joined #openstack-cinder | 16:16 | |
| *** _cjones_ has joined #openstack-cinder | 16:20 | |
| *** merooney has joined #openstack-cinder | 16:21 | |
| *** vilobhmm has joined #openstack-cinder | 16:22 | |
| *** lpetrut has quit IRC | 16:23 | |
| *** gardenshed has quit IRC | 16:25 | |
| *** crose has joined #openstack-cinder | 16:27 | |
| jms | http://paste.openstack.org/show/240231/ <-- Any ideas on _which_ dispatch.py is getting the error? I'm guessing oslo_messaging.rpc/oslo_messaging.notify ... And the way that's output, is it the ampq driver that's the _cause_ of the traceback, or is that driver output irrelevant to the traceback? | 16:28 |
| *** pboldin has joined #openstack-cinder | 16:29 | |
| pboldin | guys, please review https://review.openstack.org/182871 | 16:29 |
| *** ankit_ag has quit IRC | 16:29 | |
| *** cdelatte has joined #openstack-cinder | 16:29 | |
| *** julim has quit IRC | 16:31 | |
| e0ne | jms, pboldin: almost all are on irc meeting now | 16:31 |
| openstackgerrit | Eric Brown proposed openstack/cinder: VMware: insecure option should be exposed https://review.openstack.org/179991 | 16:32 |
| *** dannywilson has quit IRC | 16:32 | |
| *** alecv has quit IRC | 16:33 | |
| *** pboldin has quit IRC | 16:34 | |
| *** daneyon_ has quit IRC | 16:38 | |
| *** Apoorva has joined #openstack-cinder | 16:40 | |
| *** rushil has joined #openstack-cinder | 16:40 | |
| *** Apoorva has quit IRC | 16:41 | |
| *** rhedlind has joined #openstack-cinder | 16:41 | |
| asselin | nikeshm, out of curiousity, how di dyou solve your shade issue? | 16:41 |
| *** merooney has quit IRC | 16:42 | |
| openstackgerrit | Sean Chen proposed openstack/cinder: Tintri Cinder Volume driver https://review.openstack.org/185148 | 16:42 |
| *** merooney has joined #openstack-cinder | 16:42 | |
| *** merooney has quit IRC | 16:43 | |
| *** julim has joined #openstack-cinder | 16:43 | |
| *** rushil has quit IRC | 16:44 | |
| *** rushil has joined #openstack-cinder | 16:45 | |
| nikeshm | asselin: everything tried from scratc, but i think pip install shade==0.6.3 alone is enough | 16:45 |
| nikeshm | scratch | 16:45 |
| asselin__ | ok thanks | 16:45 |
| *** julim has quit IRC | 16:45 | |
| *** dulek has joined #openstack-cinder | 16:48 | |
| openstackgerrit | Sean Chen proposed openstack/cinder: Tintri Cinder Volume driver https://review.openstack.org/185148 | 16:48 |
| Swanson | xyang: Thanks! | 16:48 |
| *** annegentle has joined #openstack-cinder | 16:48 | |
| *** pboldin has joined #openstack-cinder | 16:48 | |
| *** lpetrut has joined #openstack-cinder | 16:50 | |
| *** Apoorva has joined #openstack-cinder | 16:50 | |
| nikeshm | asselin: problem is i am not getting logs/devstacklog.txt.gz file in/opt/stack/logs of jenkins slave | 16:50 |
| nikeshm | so not able to know | 16:50 |
| *** gardenshed has joined #openstack-cinder | 16:50 | |
| nikeshm | why its failing | 16:51 |
| asselin__ | nikeshm, it should be in /opt/stack/new/ | 16:51 |
| nikeshm | yes | 16:51 |
| nikeshm | sorry same | 16:51 |
| asselin__ | nikeshm, if you're using nodepool, do a nodepool hold <id> so it doesn't get deleted. then you can ssh into the vm and take a look | 16:51 |
| *** avishay has quit IRC | 16:52 | |
| *** merooney has joined #openstack-cinder | 16:52 | |
| *** rhe00 has joined #openstack-cinder | 16:53 | |
| *** rhe00 has quit IRC | 16:53 | |
| *** annegentle has quit IRC | 16:53 | |
| *** emagana has quit IRC | 16:53 | |
| *** rhedlind has quit IRC | 16:54 | |
| *** gardenshed has quit IRC | 16:57 | |
| *** sgotliv has quit IRC | 16:59 | |
| *** agarciam has quit IRC | 16:59 | |
| openstackgerrit | Sean Chen proposed openstack/cinder: Tintri Cinder Volume driver https://review.openstack.org/185148 | 16:59 |
| *** sgotliv has joined #openstack-cinder | 16:59 | |
| *** rhedlind has joined #openstack-cinder | 17:00 | |
| e0ne | fyi, our updgrade scripts https://review.openstack.org/#/c/183814/ doesn't work well:( | 17:02 |
| hemna | DuncanT, yah, I'm on the fence about it really. | 17:02 |
| nikeshm | asselin__ : http://paste.openstack.org/show/240316/ that file is no where in vm | 17:02 |
| hemna | making them work has been a PITA and unit testing them as well. | 17:02 |
| *** deepakcs has quit IRC | 17:02 | |
| hemna | but it seems that oslo.db changes will break us anyway. | 17:03 |
| nikeshm | asselin__ : like yu said i holded the vm and run test again | 17:03 |
| patrickeast | jgriffith: got a min to talk about the image cache stuff? i want to make sure i understand your latest suggestions with the periodic update of glance images and stuff | 17:03 |
| e0ne | hemna: agree | 17:03 |
| *** hemna is now known as hemnafk | 17:03 | |
| hemnafk | (beating) | 17:03 |
| patrickeast | jgriffith: i think i initially misunderstood (and probably still do) | 17:03 |
| e0ne | DuncanT: live upgrades/rollback is only one case when downgrade could help | 17:04 |
| jgriffith | patrickeast: yeah, gimmie just a minute to put some coffee on | 17:04 |
| *** tobe has joined #openstack-cinder | 17:04 | |
| openstackgerrit | xing-yang proposed openstack/cinder: Not to deduct thin volume size in capacity filter https://review.openstack.org/185764 | 17:05 |
| *** xyang has quit IRC | 17:05 | |
| *** crose has quit IRC | 17:06 | |
| *** jwcroppe has joined #openstack-cinder | 17:06 | |
| *** merooney has quit IRC | 17:07 | |
| *** merooney has joined #openstack-cinder | 17:07 | |
| Swanson | Devstack being grumpy for anyone else? | 17:07 |
| *** daneyon has joined #openstack-cinder | 17:08 | |
| *** daneyon_ has joined #openstack-cinder | 17:09 | |
| *** tobe has quit IRC | 17:09 | |
| asselin__ | Swanson, nikeshm is having issues, not sure if they're the same.... | 17:10 |
| asselin__ | ^^ | 17:10 |
| dmz | anyone using cinder in a multi-region configuratoin? I can't attach volumes since getting multi-region setup :( (it's the only thing not working in multi-region) | 17:10 |
| *** annegentle has joined #openstack-cinder | 17:10 | |
| DuncanT | e0ne: For existing migrations, I think you're right that a second, fixup migration is the only way forward... some people have already run the 'bad' migration so we end up in a place where your db depends on when you upgrade your db, which is insane | 17:11 |
| Swanson | asselin__: Oodles of role add and endpoint add failures. looks like a call changed... | 17:11 |
| *** vilobhmm has quit IRC | 17:12 | |
| *** ociuhandu has joined #openstack-cinder | 17:12 | |
| *** annegentle has quit IRC | 17:13 | |
| *** daneyon has quit IRC | 17:13 | |
| DuncanT | dmz: We use it multi-region, have been for a long time | 17:13 |
| *** ebalduf has joined #openstack-cinder | 17:13 | |
| *** agarciam has joined #openstack-cinder | 17:13 | |
| *** vilobhmm has joined #openstack-cinder | 17:13 | |
| *** markus_z has quit IRC | 17:14 | |
| *** xyang has joined #openstack-cinder | 17:14 | |
| *** julim has joined #openstack-cinder | 17:14 | |
| *** akerr has quit IRC | 17:14 | |
| xyang | winston-d, patrickeast, jgriffith: https://review.openstack.org/#/c/185764/ | 17:15 |
| *** akerr has joined #openstack-cinder | 17:15 | |
| xyang | free virtual capacity is already used for thin provisioning, so no need to add that in the capacity filter | 17:15 |
| *** vilobhmm has quit IRC | 17:16 | |
| xyang | winston-d, patrickeast, jgriffith: I just changed how the free physical capacity is used for thin | 17:16 |
| *** e0ne is now known as e0ne_ | 17:16 | |
| dulek_home | DuncanT: We want to divide the work for c-vol A/A on next's week meeting? | 17:16 |
| jgriffith | xyang: yes, perfect | 17:18 |
| jgriffith | xyang: that's what I was trying to get at | 17:18 |
| xyang | jgriffith: great, let me know your comments on the patch | 17:19 |
| jgriffith | patrickeast: shout if/when you want | 17:19 |
| dmz | DuncanT, when I added my 2nd region I have been no longer able to mount volumes; were there any cinder specific hanges needed to make it / nova understand which specific region it's working in? I've set os-region-name on command line arg & OS_REGION_NAME env variable to no success :( | 17:19 |
| patrickeast | jgriffith: i’ve got a few min now (meeting in 10) | 17:20 |
| jgriffith | patrickeast: cool, let's get going here then :) | 17:20 |
| *** e0ne_ is now known as e0ne | 17:20 | |
| patrickeast | jgriffith: so if i understand correctly your suggestion is that we have this image cache hold onto the volumes (or snapshots?) of the public snapshots, and snapshots of volume-backed instances | 17:21 |
| patrickeast | jgriffith: kind of like another way we get stuff seeded in the cache | 17:21 |
| patrickeast | jgriffith: and we could have some periodic task to update stuff in the cache from glance | 17:21 |
| jgriffith | patrickeast: yeah, so specifically, what I was envisioning was: | 17:21 |
| *** krtaylor has quit IRC | 17:22 | |
| jgriffith | patrickeast: create-volume from image ---> creates a volume template using the "image-tenant" | 17:22 |
| jgriffith | image-tenant creates volumes, then creates a "public snapshot" | 17:22 |
| jgriffith | all tenants can then access that public snapshot and do create-from snap using it | 17:22 |
| jgriffith | it's *different* | 17:23 |
| jgriffith | but same end result | 17:23 |
| patrickeast | oh so for all of the volumes backed in the cache we have a public snapshot of it available? | 17:23 |
| jgriffith | if people hate it, that's cool | 17:23 |
| jgriffith | patrickeast: right | 17:23 |
| patrickeast | that makes sense, gives people a short-cut to get fast creates if they want | 17:23 |
| jgriffith | patrickeast: then there needs to be some new internals that "check" if a backends has a snap to create the bootable volume from | 17:24 |
| jgriffith | patrickeast: right, so people can actively use the snapshot themeselves, or let the driver internals figure it out | 17:24 |
| jgriffith | eithe rway | 17:24 |
| jgriffith | either | 17:24 |
| patrickeast | gotcha | 17:24 |
| jgriffith | patrickeast: give it some thought, could be I missed some details that make it undesireable | 17:24 |
| *** akerr has quit IRC | 17:24 | |
| jgriffith | but for the most part is seemed like a useful addition to the workflow | 17:24 |
| *** e0ne has quit IRC | 17:25 | |
| jgriffith | and a bit more flexible | 17:25 |
| patrickeast | i like the idea, and do think it would be useful (i’ve been asked about public snaps by the sales guys before) but im kind of thinking maybe thats a iteration once we have the cache | 17:25 |
| jgriffith | it also gives admins some control of things | 17:25 |
| jgriffith | patrickeast: yeah, so I initially didn't like the idea of public snapshots at all | 17:25 |
| *** merooney has quit IRC | 17:25 | |
| jgriffith | patrickeast: but started asking around more, and talkes with the folks doing the AWS compat stuff some more | 17:26 |
| patrickeast | its like a feature that knows about the cache, and sits on top of it, but doesn’t really change the proposed flow of images->volumes that are re-used as needed | 17:26 |
| jgriffith | patrickeast: Im' warming up to the idea | 17:26 |
| patrickeast | yea i still don’t quite see the benefit, but there is enough demand | 17:26 |
| jgriffith | patrickeast: exactly | 17:26 |
| jgriffith | patrickeast: even better, it can change if you want, but doesn't have to | 17:26 |
| patrickeast | ok cool, so i’ll update the spec to mention this as a future use-case to make sure we track it somewhere | 17:27 |
| patrickeast | and maybe file a bp and link them | 17:27 |
| jgriffith | sweet | 17:27 |
| *** gardenshed has joined #openstack-cinder | 17:27 | |
| dmz | DuncanT, any suggestions would be most welcome :) everything google tells me is that i need the region_name passed in but i am and horizon is also having problems so it must be a config option in my api somewhere :( | 17:27 |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Updating cmd/manage.py get_arg_string() argument parser https://review.openstack.org/186094 | 17:27 |
| winston-d | jgriffith, patrickeast: I haven't read image-cache spec, so bear me for my dumb question: is image-cache suppose to be maintained by cinder and make available to all backends Cinder manages? | 17:28 |
| patrickeast | i wonder if this starts to bleed into the glance cinder backend… i heard it mentioned in the public snapshots session, but i don’t really know exactly what the scope is for it | 17:28 |
| patrickeast | winston-d: yea the idea is you can enable it with a backend config option and anyone can use it | 17:28 |
| patrickeast | winston-d: it only makes sense for some backends, so it would require some docs to suggestion when its best to enable | 17:28 |
| winston-d | patrickeast: do you have the spec link handy? | 17:29 |
| patrickeast | winston-d: https://review.openstack.org/#/c/182520/ | 17:29 |
| winston-d | patrickeast: perfect, thx | 17:29 |
| *** gardenshed has quit IRC | 17:30 | |
| *** ebalduf has quit IRC | 17:30 | |
| *** emagana has joined #openstack-cinder | 17:31 | |
| jgriffith | winston-d: if you haven't seen it, it's based loosely off of this: https://goo.gl/HWtizZ | 17:32 |
| *** sgotliv has quit IRC | 17:32 | |
| *** crose has joined #openstack-cinder | 17:33 | |
| *** ebalduf has joined #openstack-cinder | 17:33 | |
| winston-d | jgriffith: thx | 17:33 |
| dulek_home | jgriffith: Do you had time to take a look at TaskFlow refactoring? | 17:33 |
| *** crose has quit IRC | 17:34 | |
| dulek_home | jgriffith: To me it seems that there's at least two separate issues - too much logging and the fact that code is harder to read. | 17:35 |
| jgriffith | dulek_home: sadly no, I took the week-end off and did "not much at all" | 17:35 |
| *** aix has quit IRC | 17:35 | |
| jgriffith | dulek_home: the logging doesn't bother me as much as it does some folks | 17:35 |
| dulek_home | jgriffith: Similar stuff from my side, jetlag killed my productivity. ;) | 17:35 |
| jgriffith | dulek_home: I can use grep pretty easily :) | 17:35 |
| *** julim has quit IRC | 17:35 | |
| *** vilobhmm has joined #openstack-cinder | 17:35 | |
| jgriffith | dulek_home: the complexity and following the code is the thing for me | 17:35 |
| jgriffith | dulek_home: hehe.. yeah, where is home? | 17:36 |
| dulek_home | jgriffith: Poland, from UTC-7 to UTC+2. ;) | 17:37 |
| *** crose has joined #openstack-cinder | 17:37 | |
| nikeshm | asselin: got the issue,my vm ip is 10.0.1.x and FIXED_RANGE=${DEVSTACK_GATE_FIXED_RANGE:-10.1.0.0/20} in devstack-gate/devstack-vm-gate.sh | 17:37 |
| nikeshm | sorry my vm ip is 10.1.0.x | 17:37 |
| asselin__ | nikeshm, ok cool | 17:37 |
| dulek_home | jgriffith: I can work with harlowja_ to get logging right and start for example with refactoring easier flows - like these in c-api and c-sch. | 17:37 |
| harlowja_ | dulek_home cool, sputnix (another taskflow core) has some ideas (and wants something similar there to) | 17:38 |
| winston-d | jgriffith, dulek_home: both log and 'hard to read' part bother me | 17:38 |
| harlowja_ | make exceptions betteer (more looking like tracebacks) | 17:38 |
| dulek_home | jgriffith: It would be great if you could take a look on c-vol one. I haven't seen this code before TaskFlow got in, so you're way more familiar. | 17:38 |
| *** vilobhmm has quit IRC | 17:38 | |
| harlowja_ | dulek_home https://review.openstack.org/#/c/184814/ | 17:38 |
| *** julim has joined #openstack-cinder | 17:39 | |
| harlowja_ | and for all its worth, if the initial code in cinder can be made better (easier to read...), let's do it (it was after all made a year + ago) | 17:39 |
| *** juzuluag has quit IRC | 17:39 | |
| *** hemnafk is now known as hemna | 17:39 | |
| dulek_home | harlowja_: What actually sputnix want to change? | 17:40 |
| dulek_home | harlowja_: You mean exceptions? | 17:40 |
| harlowja_ | dulek_home soooo, when a failure occurs, add a way to make the failure string representation include various engine state information (what has executed, what the active results were...) | 17:40 |
| harlowja_ | this makes it look more like a stack trace | 17:41 |
| dulek_home | harlowja_: Ah, that's cool thing. Now the exception is likely to be displayed twice. | 17:41 |
| harlowja_ | possibly, not sure what cinder is doing with it :-P | 17:41 |
| harlowja_ | dulek_home http://docs.openstack.org/developer/taskflow/types.html#module-taskflow.types.failure just needs to have a reference to the engine it came from (and its __str__ or other method needs to extract info from it) | 17:42 |
| harlowja_ | probably http://docs.openstack.org/developer/taskflow/types.html#taskflow.types.failure.Failure.pformat (that method?) needs to do this | 17:42 |
| harlowja_ | something along that line | 17:42 |
| dulek_home | Okay | 17:43 |
| nikeshm | asselin: thanks,my provider has a flavor of ram 16gb and same i am giving in nodepool and in dashboard also its showing that nodepool vm have a ram 16 gb but after loging it showing only 6 gb ram | 17:43 |
| harlowja_ | dulek_home right now cinder has https://github.com/openstack/cinder/blob/master/cinder/flow_utils.py#L48 (which it could also do things with to make similar information) | 17:43 |
| harlowja_ | that things is whats attaching into the engine state transitions and logging failures and stuff | 17:44 |
| dulek_home | harlowja_: Yeah, I'm aware of that stuff :) | 17:44 |
| harlowja_ | *so cinder code could format the stuff as it wants already | 17:44 |
| *** dustins has quit IRC | 17:45 | |
| jms | http://paste.openstack.org/show/240231/ <-- Any ideas on _which_ dispatch.py is getting the error? I'm guessing oslo_messaging.rpc/oslo_messaging.notify ... And the way that's output, is it the ampq driver that's the _cause_ of the traceback, or is that driver output irrelevant to the traceback? | 17:45 |
| asselin__ | nikeshm, not sure.... | 17:45 |
| asselin__ | nikeshm, signing off...will be back later | 17:46 |
| *** jordanP has quit IRC | 17:46 | |
| winston-d | jms: before digging into that, how's your debugging so far? | 17:46 |
| dulek_home | jgriffith: So do you agree on the approach to the refactoring? | 17:47 |
| *** earlephilhower has joined #openstack-cinder | 17:47 | |
| nikeshm | asselin:ok | 17:47 |
| jgriffith | dulek_home: kinda | 17:47 |
| dulek_home | jgriffith: What's bugging you? | 17:48 |
| jgriffith | dulek_home: the thing is taskflow doesn't really do anything with c-vol | 17:48 |
| jgriffith | dulek_home: it replaced code in c-api and c-manager | 17:48 |
| jgriffith | dulek_home: so what you proposed means I don't have to do anything :) | 17:48 |
| jgriffith | dulek_home: not that there's anything wrong with that :) | 17:48 |
| jms | winston-d: Still the same.... the controller just doesn't look for the backends on the second node. That trace is the only thing I'm seeing... not seeing any rabbit communication errors, no auth errors, etc... :/ | 17:48 |
| *** vilobhmm has joined #openstack-cinder | 17:48 | |
| *** vilobhmm has quit IRC | 17:48 | |
| *** merooney has joined #openstack-cinder | 17:49 | |
| *** vilobhmm has joined #openstack-cinder | 17:49 | |
| dulek_home | jgriffith: https://github.com/openstack/cinder/blob/master/cinder/volume/flows/manager/create_volume.py | 17:49 |
| dulek_home | jgriffith: This is what I've meant. | 17:49 |
| jgriffith | dulek_home: so that's the only call we really use taskflow for currently | 17:49 |
| jgriffith | dulek_home: yeah, sure | 17:50 |
| jgriffith | that sounds good | 17:50 |
| jms | winston-d: Now I could be absolutly looking over some obvious thing ... :/ | 17:50 |
| *** emagana has quit IRC | 17:50 | |
| dulek_home | jgriffith: API one also has ~850 lines of code. | 17:50 |
| *** merooney has quit IRC | 17:51 | |
| *** asselin__ has quit IRC | 17:51 | |
| winston-d | jms: did you check c-sch log and find the zfs backend reporting stats? | 17:51 |
| harlowja_ | dulek_home https://review.openstack.org/#/c/185116/ could also make things easier/more elegant(?) for cinder to | 17:52 |
| harlowja_ | see example in there | 17:52 |
| *** akerr has joined #openstack-cinder | 17:52 | |
| dulek_home | jgriffith: Just to clarify for myself - what's the difference between c-manager and c-vol? There's c-vol in DevStack... | 17:52 |
| *** ebalduf has quit IRC | 17:52 | |
| *** dannywilson has joined #openstack-cinder | 17:52 | |
| *** chlong has quit IRC | 17:53 | |
| *** vilobhmm has quit IRC | 17:53 | |
| jgriffith | dulek_home: so theyr'e kinda the same | 17:54 |
| jgriffith | dulek_home: if you look in the code though... | 17:54 |
| dulek_home | harlowja_: I don't think that simple building of the flow can be made easier. ;) | 17:54 |
| jgriffith | dulek_home: manager.py actually coordinates calls to the volume driver | 17:54 |
| jgriffith | dulek_home: if you ask guitarzan it's pointless :) | 17:54 |
| *** dannywil_ has joined #openstack-cinder | 17:55 | |
| harlowja_ | dulek_home hmmm, idk, maybe u are right :-P | 17:55 |
| jgriffith | harlowja_: dulek_home one thing I was looking at was seperating things a bit | 17:56 |
| *** dannywilson has quit IRC | 17:56 | |
| jgriffith | harlowja_: dulek_home building the flow, vs running the flow | 17:56 |
| jgriffith | harlowja_: I think that might help? | 17:56 |
| harlowja_ | isn't that already sorta split? | 17:56 |
| jgriffith | IDK | 17:56 |
| jgriffith | harlowja_: yes | 17:56 |
| jgriffith | harlowja_: you actually did a bunch of work on that for me once already :) | 17:56 |
| jgriffith | harlowja_: dulek_home I'll spend some time looking at it | 17:57 |
| harlowja_ | possibly, my memory is getting bad, lol | 17:57 |
| harlowja_ | getting old and all | 17:57 |
| harlowja_ | lol | 17:57 |
| jgriffith | harlowja_: hehe | 17:57 |
| jgriffith | welcome my friend! | 17:57 |
| harlowja_ | :-P | 17:57 |
| harlowja_ | oh hi, who are u | 17:57 |
| harlowja_ | how did i end up here, lol | 17:57 |
| dulek_home | :D | 17:57 |
| *** merooney has joined #openstack-cinder | 17:58 | |
| jms | winston-d: Not on the controller sched. When I had the sched running on the 2nd host I was getting both (in API log anyway), but not since I disabled it. On the 2nd host I was seeing both LVM and the zfs backend in sched when I had sched running on it. | 17:58 |
| * dulek_home Still trying to get c-vol and c-manager things right. | 17:58 | |
| dulek_home | jgriffith: So what you mean is c-vol is volume driver and c-manager is the actual service running? | 17:59 |
| jgriffith | dulek_home: rpc calls get routed to the manger which then routes to the actual driver | 17:59 |
| jgriffith | dulek_home: the combination fo manager and driver make up the c-vol service | 17:59 |
| *** rmesta has quit IRC | 17:59 | |
| jgriffith | dulek_home: the driver code is "only" interested in accessing the device | 18:00 |
| *** rmesta has joined #openstack-cinder | 18:00 | |
| jgriffith | dulek_home: the manager is a bridge between the device-driver and the rpc layer | 18:00 |
| *** pradipta has quit IRC | 18:00 | |
| *** rushiagr_away is now known as rushiagr | 18:00 | |
| dulek_home | jgriffith: Okay, get it. So actually the flow is running in manager (which is part of c-vol). | 18:00 |
| dulek_home | jgriffith: I get the difference though. | 18:01 |
| winston-d | jms: so basically nothing (lvm/zfs) on your scheduler log on 1st host? | 18:01 |
| *** IanGovett has quit IRC | 18:01 | |
| jgriffith | dulek_home: right | 18:02 |
| *** dims_ has quit IRC | 18:02 | |
| *** dims_ has joined #openstack-cinder | 18:03 | |
| dulek_home | jgriffith: Thanks for explaining. Okay, gotta go. Have a nice day everyone! | 18:03 |
| *** merooney has quit IRC | 18:03 | |
| *** dulek_home has quit IRC | 18:03 | |
| *** manishg has joined #openstack-cinder | 18:05 | |
| *** e0ne has joined #openstack-cinder | 18:06 | |
| *** e0ne is now known as e0ne_ | 18:06 | |
| *** e0ne_ is now known as e0ne | 18:07 | |
| jms | winston-d: lvm is. The one that's defined in that hosts cinder.conf | 18:10 |
| *** merooney has joined #openstack-cinder | 18:10 | |
| *** setmason has quit IRC | 18:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/186112 | 18:11 |
| *** merooney has quit IRC | 18:12 | |
| *** setmason has joined #openstack-cinder | 18:13 | |
| winston-d | jms: is there problem for your 2nd node to connect to your rabbit/mysql (which I suppose are on 1st host)? | 18:15 |
| jms | winston-d: Nope, connects fine. Haven't seen any sort of errors from rabbit, a failed mysql connection, etc... | 18:16 |
| *** krtaylor has joined #openstack-cinder | 18:16 | |
| *** annashen has joined #openstack-cinder | 18:17 | |
| jms | winston-d: When I had the scheduler running on the zfs node, I was able to create a volume (cinder create), *if* I created it on that node... but going through the controller still failed. | 18:17 |
| winston-d | jms: then we can't explain why no stats update can be seen on scheduler. | 18:17 |
| guitarzan | jgriffith: hah, hey, I use c-vol too :) | 18:17 |
| winston-d | jms: can you post the output of 'keystone endpoint-list' from both 1st & 2nd node? | 18:18 |
| winston-d | jms: when you create from the controller, what error did you see? vol stuck at 'creating' or 'error'? | 18:19 |
| jms | winston-d: http://paste.openstack.org/show/240423/ | 18:20 |
| *** BharatK has quit IRC | 18:20 | |
| *** setmason has quit IRC | 18:21 | |
| *** kmartin has quit IRC | 18:22 | |
| jgriffith | guitarzan: hehe.. but not because you want to :) | 18:23 |
| *** vilobhmm has joined #openstack-cinder | 18:23 | |
| winston-d | jms: this is for 1st node? | 18:23 |
| jms | winston-d: Error ... the error is: No weighted hosts found ... It's trying on the LVM storage, but fails because the volume_backend_name extra-spec doesn't match | 18:23 |
| *** merooney has joined #openstack-cinder | 18:23 | |
| guitarzan | jgriffith: hmm, I'm not sure if that's true or not... might be | 18:23 |
| jgriffith | guitarzan: LOL... I could be confused | 18:24 |
| winston-d | jms: who is trying? the scheduler on first or 2nd node? | 18:24 |
| jms | winston-d: That is ran from the 2nd node. | 18:24 |
| guitarzan | it does make some decisions that make our lives a little miserable | 18:24 |
| guitarzan | but it doesn't typically do much harm :) | 18:24 |
| jgriffith | guitarzan: the conversation is coming back to me now :) | 18:24 |
| jms | winston-d: The 1st node is trying. That's the only node that has a scheduler running (after I stopped it from the 2nd node, from my understanding that there should be a single one yesterday). | 18:25 |
| *** ankit_ag has joined #openstack-cinder | 18:25 | |
| *** merooney has quit IRC | 18:26 | |
| *** fthiagogv has joined #openstack-cinder | 18:26 | |
| *** vilobhmm1 has joined #openstack-cinder | 18:27 | |
| jms | winston-d: But even when I did have the sched running on the zfs host, it failed creation unless I ran the 'cinder create' on the zfs host. Would get the no host error if I ran it from the controller cli, or through horizon (I think that's the name of that web thing). | 18:27 |
| winston-d | xyang: could you explain why free capacity is only deduct when backend doesn't suport thin? this logic wasn't there and doesn't seem related to the bug. | 18:29 |
| winston-d | xyang: ok, i just noticed that you left comment on previous change. looking | 18:30 |
| xyang | winston-d: hi, ok | 18:30 |
| *** ankit_ag has quit IRC | 18:30 | |
| *** vilobhmm has quit IRC | 18:30 | |
| *** merooney has joined #openstack-cinder | 18:31 | |
| *** setmason has joined #openstack-cinder | 18:31 | |
| openstackgerrit | Alfredo Moralejo proposed openstack/cinder: rbd driver in cinder does not manage glance images multi-location https://review.openstack.org/182616 | 18:32 |
| winston-d | xyang: i don't agree with your argument about free capacity for thin being deduct (reserved) from schedule would cause any signifincant problem. | 18:32 |
| *** merooney has quit IRC | 18:33 | |
| winston-d | xyang: as you mentioned in your reply to wang hao's comment, schedule stats are being updated every 60 secs. | 18:33 |
| xyang | winston-d: since free_capacity is for physical capacity, why do we need to deduct the new volume size from it? that's why I think they are related | 18:33 |
| xyang | winston-d: I guess I don't understand why we deduct the volume size there | 18:33 |
| *** BharatK has joined #openstack-cinder | 18:33 | |
| *** dulek has quit IRC | 18:33 | |
| *** dustins has joined #openstack-cinder | 18:33 | |
| xyang | winston-d: is the free_capacity we get in capacity_filter a result after the deduction? | 18:34 |
| winston-d | xyang: if you don't, basically you are treat thin backend as inifinte for 60 secs. | 18:34 |
| *** setmason has quit IRC | 18:34 | |
| winston-d | which is totally wrong | 18:34 |
| *** kmartin has joined #openstack-cinder | 18:34 | |
| *** setmason has joined #openstack-cinder | 18:34 | |
| winston-d | imgine that you have 200 vol creates coming in in 60secs, before your backend is able to send updates to sch. | 18:35 |
| xyang | winston-d: if a new thin volume is 10GB and not written yet, its physical capacity is 0 | 18:35 |
| guitarzan | winston-d: thin is infinite! :) | 18:36 |
| winston-d | xyang: so you can create as many vols as you want, that's no right. if that's the case, why do we bother to do oversubscriptioon ratio at all? | 18:36 |
| xyang | winston-d: I'd like patrickeast to take a look too. It seems that we should use the same standard in both places | 18:36 |
| hemna | winston-d, doesn't every driver basically have that problem ? | 18:37 |
| winston-d | these two are totally unrelated issue. | 18:37 |
| *** rushil has quit IRC | 18:37 | |
| winston-d | if you want to fix that, please create a separate bug | 18:37 |
| hemna | as our driver doesn't keep track of # of volumes created between get_volume_stats() updates | 18:37 |
| winston-d | hemna: that's why i want scheduler to be cautious | 18:37 |
| patrickeast | winston-d: xyang: you guys are talking about the changes in host_manager, right? | 18:37 |
| xyang | winston-d: so during this 60 seconds, we assume the new thin volume space will be totally written | 18:37 |
| hemna | and like xyang says, if each of those volumes created doesn't get data, then the reporting from the array will be the same as prior to their creation | 18:38 |
| xyang | winston-d: then in capacity filter, we think a new thin volume will not be written. that is where my confusion is coming from | 18:38 |
| hemna | we might have allocated #'s though, not sure | 18:38 |
| patrickeast | so we either play it safe and assume they might be full, or assume they are empty and take up what amounts to 0 space | 18:38 |
| patrickeast | either way its not super great | 18:38 |
| winston-d | xyang: if you want, you can do a more complex math, i.e. deduct the size/oversus_ratio | 18:38 |
| winston-d | xyang: but no deduct at all will create problem. | 18:38 |
| patrickeast | i’m ok with deducting the full amount there | 18:39 |
| patrickeast | as long as we keep the other changes in capacity_filter.py | 18:39 |
| patrickeast | they are two different issues in my mind | 18:39 |
| xyang | winston-d: I guess what I don't understand is that, if we deduct in host_manager, why don't we consider the size in capacity_filter? | 18:39 |
| winston-d | xyang: you can have your QA team to design a case for this and see how your backend works. | 18:39 |
| winston-d | patrickeast: yes, two different issues. | 18:40 |
| xyang | winston-d: will this (free_capacity-volume_size) be used in capacity_filter? | 18:40 |
| winston-d | xyang: we do in capacity filer, i don't understand your argument, TBH | 18:40 |
| *** gardenshed has joined #openstack-cinder | 18:40 | |
| winston-d | xyang: it will, for next create request | 18:40 |
| jgriffith | winston-d: FWIW I think you're describing exactly what I was trying to propose | 18:41 |
| asselin | nikeshm, btw, your ram issue is not likely related to nodepool. Take a look at nova and see what's going on there | 18:41 |
| *** akerr has quit IRC | 18:43 | |
| xyang | winston-d: are you okay with the changes in capacity_filter only? | 18:43 |
| *** gardenshed has quit IRC | 18:44 | |
| *** akerr has joined #openstack-cinder | 18:45 | |
| jungleboyj | jgriffith: Do you know if it is still possible to run into race conditions in c-api if running active active? | 18:45 |
| jungleboyj | https://etherpad.openstack.org/p/kilo-crossproject-ha-integration | 18:45 |
| e0ne | jungleboyj: i've got race in c-api in Juno | 18:46 |
| openstackgerrit | Rushil Chugh proposed openstack/cinder: Avoid LUN ID collisions in NetApp iSCSI drivers https://review.openstack.org/179239 | 18:47 |
| *** ebalduf has joined #openstack-cinder | 18:47 | |
| *** merooney has joined #openstack-cinder | 18:47 | |
| *** rhedlind has quit IRC | 18:47 | |
| jungleboyj | e0ne: Ok, and I don't think we did anything in Kilo to change that either. | 18:47 |
| e0ne | jungleboyj: agree | 18:47 |
| *** vilobhmm has joined #openstack-cinder | 18:47 | |
| *** vilobhmm1 has quit IRC | 18:48 | |
| *** rushil has joined #openstack-cinder | 18:48 | |
| jgriffith | jungleboyj: well *someobdy* added a comment that "this it valid issue in Juno" so apparantly it's a problem | 18:48 |
| jgriffith | jungleboyj: note they said "this it" not me :) | 18:48 |
| *** BharatK has quit IRC | 18:48 | |
| e0ne | jgriffith: maybe it was me | 18:49 |
| jgriffith | e0ne: darn you! :) | 18:49 |
| jungleboyj | jgriffith: So, you would agree that we still have issues there? | 18:50 |
| e0ne | fixed to "this is" | 18:50 |
| *** merooney has quit IRC | 18:50 | |
| jungleboyj | :-) | 18:50 |
| jgriffith | jungleboyj: not saying that :) | 18:50 |
| jgriffith | jungleboyj: so that's an odd one to call out IMHO | 18:51 |
| *** merooney has joined #openstack-cinder | 18:51 | |
| jgriffith | jungleboyj: we have all sorts of issues in other places I think (races) | 18:51 |
| jgriffith | jungleboyj: but I think DuncanT has been the most involved/vocal on that topic | 18:51 |
| e0ne | https://etherpad.openstack.org/p/cinder-active-active-vol-service-issues | 18:51 |
| * jungleboyj missed that tag on the end of the line. Sorry about that. | 18:52 | |
| winston-d | xyang: I've left new comment. | 18:52 |
| *** tobe has joined #openstack-cinder | 18:53 | |
| xyang | winston-d: thanks | 18:54 |
| *** leeantho has quit IRC | 18:55 | |
| xyang | winston: I am ok just to make changes in capacity_filter now. then both patrickeast and I can do more testing to see if anything else needs to be changed | 18:55 |
| patrickeast | sounds good to me | 18:56 |
| winston-d | xyang, patrickeast: thx | 18:58 |
| *** tobe has quit IRC | 18:58 | |
| *** nlevinki has joined #openstack-cinder | 18:58 | |
| *** e0ne has quit IRC | 18:59 | |
| winston-d | jms: so, i'm pretty sure there is something wrong with your setup. | 19:00 |
| *** ik__ has joined #openstack-cinder | 19:00 | |
| ik__ | any idea why this would happen? http://paste.openstack.org/raw/240475/ | 19:00 |
| winston-d | jms: e.g. your c-vol service on 2nd isn't able to send updates to scheduler on 1st node. | 19:00 |
| *** rushiagr is now known as rushiagr_away | 19:01 | |
| *** bandwidth has joined #openstack-cinder | 19:01 | |
| bandwidth | I have a hard time configuring notification_topics for cinder (kilo) | 19:02 |
| bandwidth | it was working fine before Kilo, but now, the bindings aren't configured properly in rabbitmq | 19:02 |
| bandwidth | someone can help? | 19:02 |
| *** merooney has quit IRC | 19:03 | |
| *** ik__ has left #openstack-cinder | 19:05 | |
| winston-d | bandwidth: what exact problem are you seeing with notification_topics with kilo cinder? | 19:07 |
| *** eharney has quit IRC | 19:08 | |
| vilobhmm | scottda : ping | 19:08 |
| scottda | :Hi vilobhmm | 19:10 |
| vilobhmm | when you see that you can't detach volumes as they are stuck in "detaching" state how do you get around this problem….how do you make sure that the data on those volumes is accessible (as we can't attach the volume to any other instance) | 19:10 |
| nikeshm | asselin: tried to create a cirros image with flavor of 8 gb,ssh to cirros vm,ram is 8 gb but same thing when tried with nodepool vm its 6gb,you are correct better to check nova | 19:10 |
| vilobhmm | since the state of volumes in detaching ( and not available) | 19:10 |
| scottda | You have to see the state of the Volume(s) in Nova (is it attached to an instance), in Nova BlockDeviceMapping table (if BDM table says it's attached, you won't be able to attach it elsewhere) .... | 19:11 |
| scottda | State in Cinder DB (needs to be 'available' to re-attach).. | 19:11 |
| scottda | and state on storage backend (needs a Cinder terminate_connection or manual un-export of volume to compute host. | 19:12 |
| scottda | In other words, solving the problem is a big mess | 19:12 |
| *** annashen has quit IRC | 19:14 | |
| *** eharney has joined #openstack-cinder | 19:15 | |
| xyang | winston-d: are you still there? | 19:16 |
| scottda | vilobhmm: Some notes and links here: https://etherpad.openstack.org/p/CinderReset-stateAndForceDelete | 19:17 |
| *** merooney has joined #openstack-cinder | 19:17 | |
| winston-d | xyang: yes | 19:17 |
| xyang | winston-d: line 85 is not covered. | 19:18 |
| xyang | winston-d: https://review.openstack.org/#/c/185764/3/cinder/scheduler/filters/capacity_filter.py | 19:18 |
| vilobhmm | scottda : I agree…as DuncanT , myself and all of us saw in active active volume services design summit talk….just wanted to know if there was a work around rather that doing the db route | 19:18 |
| vilobhmm | thanks for the link | 19:18 |
| xyang | winston-d: the case for thin and over subscription >=1, do we want to check if free capacity is 0 | 19:19 |
| nikeshm | asselin : this i found in nova for nodepool vm http://paste.openstack.org/show/240527/ | 19:20 |
| *** kmartin has quit IRC | 19:20 | |
| winston-d | xyang: i don't think we need an extra check, because it's already covered here: https://github.com/openstack/cinder/blob/master/cinder/scheduler/filters/capacity_filter.py#L114 | 19:22 |
| xyang | winston-d: that is using free_virtual_capacity. do we also want to check free_physical_capacity is 0? | 19:23 |
| xyang | winston-d: if free physical capacity, do we still allow thin provisioning to happen | 19:23 |
| xyang | winston-d: if free physical capacity is 0 | 19:24 |
| winston-d | xyang: nope, free_virtual = free * host_state.max_over_subscription_ratio | 19:24 |
| winston-d | xyang: free virtual will be 0 if free is 0 | 19:24 |
| winston-d | xyang: and free_virtual >= volume_size will be False | 19:24 |
| xyang | winston-d: that is a good point | 19:25 |
| winston-d | make sense? | 19:25 |
| xyang | winston-d: yes | 19:25 |
| scottda | vilobhmm: I'm working on some POC code to fix, but it's a few days out. | 19:25 |
| winston-d | scottda: actually, if a volume is marked as 'available' in cinder but still in Nova BDM, no one will be able to stop user from attaching the volume and other instances. | 19:26 |
| vilobhmm | let me know if any help is needed | 19:26 |
| vilobhmm | we are seeing this internally as well on a PoC and would like to help you so that it benefits everyone | 19:27 |
| *** theanalyst has quit IRC | 19:27 | |
| scottda | winston-d: I think Nova will stop the user from attaching again, since it searches the BDM table and will fail if it finds an entry for the volume. | 19:29 |
| *** theanalyst has joined #openstack-cinder | 19:30 | |
| vilobhmm | winston-d, scottda : I dont see status field in nova.bdm http://paste.openstack.org/show/240528/ | 19:30 |
| vilobhmm | scottda : I agree what you said | 19:30 |
| *** ebalduf has quit IRC | 19:32 | |
| *** madskier has joined #openstack-cinder | 19:32 | |
| *** vokt has quit IRC | 19:34 | |
| winston-d | scottda, vilobhmm: can you point out where does Nova do the check? I couldn't find it after going through API extension, Compute Api and Compute Manager. | 19:35 |
| openstackgerrit | xing-yang proposed openstack/cinder: Not to deduct thin volume size in capacity filter https://review.openstack.org/185764 | 19:35 |
| *** xyang has quit IRC | 19:35 | |
| vilobhmm | https://github.com/openstack/nova/blob/master/nova/volume/cinder.py#L309 | 19:36 |
| *** kvidvans has quit IRC | 19:36 | |
| *** jgriffith has quit IRC | 19:36 | |
| vilobhmm | winston-d, scottda : ^^ | 19:36 |
| *** jgriffith has joined #openstack-cinder | 19:36 | |
| winston-d | vilobhmm: what I said was: if a volume is marked as 'available' in cinder but still in Nova BDM, no one will be able to stop user from attaching the volume and other instances. | 19:36 |
| winston-d | vilobhmm: checking volume's status doesn't gate the case I mentioned | 19:37 |
| scottda | Off to a meeting just now, I'll look in the code for where Nova checks BDM mapping. I'm pretty sure it's there. | 19:37 |
| winston-d | scottda: ok. | 19:38 |
| vilobhmm | winston-d : what you said makes sense….may be I read betwn the lines..missed .."but still in Nova BDM" | 19:38 |
| vilobhmm | let me point you to bdm code | 19:38 |
| *** lpabon has joined #openstack-cinder | 19:41 | |
| *** nlevinki has quit IRC | 19:42 | |
| *** madskier has quit IRC | 19:47 | |
| *** ebalduf has joined #openstack-cinder | 19:48 | |
| *** ebalduf has quit IRC | 19:52 | |
| *** ebalduf has joined #openstack-cinder | 19:54 | |
| *** ociuhandu has quit IRC | 19:55 | |
| *** ganso_ has joined #openstack-cinder | 19:56 | |
| *** vokt has joined #openstack-cinder | 19:57 | |
| *** jungleboyj has quit IRC | 19:58 | |
| *** juzuluag has joined #openstack-cinder | 19:58 | |
| *** xyang has joined #openstack-cinder | 19:59 | |
| *** sgotliv has joined #openstack-cinder | 19:59 | |
| *** kmartin has joined #openstack-cinder | 20:00 | |
| *** leeantho has joined #openstack-cinder | 20:00 | |
| *** Lee1092 has quit IRC | 20:01 | |
| *** Rockyg has joined #openstack-cinder | 20:01 | |
| *** juzuluag has quit IRC | 20:03 | |
| *** simondodsley has quit IRC | 20:03 | |
| nikeshm | asselin:i tried to manually create vm with nodepool image,but still its taking 6 gb,and in nova i have around 90GB,is any kernel limit of nodepool image | 20:03 |
| asselin | nikeshm, I don't know.... | 20:04 |
| nikeshm | ok | 20:05 |
| nikeshm | let me ask in infra | 20:05 |
| vilobhmm | winston-d : https://github.com/openstack/nova/blob/a8259c4d8dd475f993eef75b268e952a0bafc341/nova/api/openstack/compute/plugins/v3/block_device_mapping.py#L65; https://github.com/openstack/nova/blob/master/nova/block_device.py#L175 ; | 20:05 |
| openstackgerrit | Vilobh Meshram proposed openstack/cinder: Nested Quota : Create allocated column in cinder.quotas https://review.openstack.org/185704 | 20:12 |
| bandwidth | winston-d: normally, cinder should create bindings on the configured exchange to route the message based on the routing-key | 20:12 |
| bandwidth | winston-d: i completely removed the exchange (cinder), restarted the service, it recreated the exchange with the bindings (cinder-volumes, cinder-backups, etc) but without the configured notification_topics provided (notification_topics=notifications,xyx...) | 20:13 |
| *** annashen has joined #openstack-cinder | 20:15 | |
| *** IanGovett has joined #openstack-cinder | 20:16 | |
| *** timcl1 has quit IRC | 20:17 | |
| *** esker has quit IRC | 20:19 | |
| jms | winston-d: Yeah, I just don't know _what_ is being blocked... I've disabled the firewall for now, so it isn't the port being blocked. As I mentioned, I haven't seen any errors anywhere about failed connectes, it's showing it's connecting to the amqp server, the database connection seems to be valid (and I can connect with mysql cli client from zfs host to db on controller), api-paste.ini sections appear fine. :/ | 20:19 |
| *** madskier has joined #openstack-cinder | 20:21 | |
| *** ebalduf has quit IRC | 20:21 | |
| jms | winston-d: c-volume.log has output of: "Notifying Schedulers of capabilities ... _publish_service_capabilities" ... So, I'm guessing it may be trying, or thinking about, updating. But I don't know of a way to increase verbosity to see what it's doing more than having debug/verbose true, or using pdb to try tracing cinder-volume :/ | 20:28 |
| scottda | winston-d: You are right. There is nothing in Nova that prevents attaching a volume to 2 instances at once, if you have used reset-state to make the volume 'available'. | 20:32 |
| *** lpabon has quit IRC | 20:32 | |
| scottda | Looks like we didn't need all that complicated code for multi-attach :) | 20:32 |
| *** julim has quit IRC | 20:34 | |
| *** barra204 has joined #openstack-cinder | 20:35 | |
| *** annegentle has joined #openstack-cinder | 20:37 | |
| *** jungleboyj has joined #openstack-cinder | 20:37 | |
| hemna | lol | 20:39 |
| *** rushil has quit IRC | 20:40 | |
| *** IanGovett has quit IRC | 20:41 | |
| *** tobe has joined #openstack-cinder | 20:42 | |
| *** rushil has joined #openstack-cinder | 20:44 | |
| *** esker has joined #openstack-cinder | 20:46 | |
| *** tobe has quit IRC | 20:46 | |
| *** annashen has quit IRC | 20:47 | |
| *** nkrinner has quit IRC | 20:48 | |
| *** ronis has quit IRC | 20:49 | |
| *** barra204 has quit IRC | 20:50 | |
| *** bswartz has quit IRC | 20:50 | |
| winston-d | scottda: yeah, that's what we've observed in our production env. we do hit a lot of issues with attach/detach. it's really a PITA. | 20:52 |
| winston-d | inconsistency between Cinder and Nova; inconsistency between Nova Api cell and child cell. | 20:53 |
| *** rushil has quit IRC | 20:53 | |
| winston-d | jms: cinder-volume publish its stats via a fanout call, so basically just send the message and don't care who receives it. | 20:54 |
| winston-d | jms: you may want to check your rabbit queues, see if the fanout queue is there. you can even dump messages from there. | 20:54 |
| *** rasoto_ has quit IRC | 20:55 | |
| *** rushil has joined #openstack-cinder | 21:00 | |
| *** daneyon_ has quit IRC | 21:00 | |
| *** ociuhandu has joined #openstack-cinder | 21:01 | |
| *** esker has quit IRC | 21:01 | |
| *** ociuhandu has quit IRC | 21:02 | |
| *** dustins has quit IRC | 21:02 | |
| *** rushil has quit IRC | 21:04 | |
| *** ociuhandu has joined #openstack-cinder | 21:04 | |
| *** rushil has joined #openstack-cinder | 21:04 | |
| *** akerr has quit IRC | 21:05 | |
| *** ebalduf has joined #openstack-cinder | 21:06 | |
| *** rushil has quit IRC | 21:06 | |
| *** madskier has quit IRC | 21:07 | |
| openstackgerrit | Jon Bernard proposed openstack/os-brick: Add RBD connector https://review.openstack.org/186172 | 21:11 |
| anteaya | so I was just going through the meeting backscroll | 21:12 |
| anteaya | and on the top of db downgrades | 21:12 |
| anteaya | after the ops tagging working group session I chatted with a few ops before lunch | 21:13 |
| *** bkopilov_wfh has quit IRC | 21:13 | |
| anteaya | they never use a db downgrade | 21:13 |
| anteaya | since they are never sure it will work | 21:13 |
| anteaya | Jon Proulx was one of the people I was talking with | 21:14 |
| anteaya | if anyone wants to reach out to him and get his firsthand thoughts on the matter | 21:14 |
| *** bswartz has joined #openstack-cinder | 21:18 | |
| *** ebalduf has quit IRC | 21:18 | |
| *** Yogi11 has quit IRC | 21:19 | |
| *** merooney has quit IRC | 21:19 | |
| openstackgerrit | Jon Bernard proposed openstack/os-brick: Add RBD connector https://review.openstack.org/186172 | 21:21 |
| *** bkopilov has joined #openstack-cinder | 21:23 | |
| *** dannywil_ has quit IRC | 21:27 | |
| *** dannywilson has joined #openstack-cinder | 21:27 | |
| *** akshai has quit IRC | 21:31 | |
| *** bkopilov has quit IRC | 21:31 | |
| *** merooney has joined #openstack-cinder | 21:32 | |
| openstackgerrit | Tom Barron proposed openstack/cinder: Add standard QoS spec support to cDOT drivers https://review.openstack.org/179352 | 21:40 |
| *** merooney has quit IRC | 21:40 | |
| jms | winston-d: Hrmm... Thanks for the rabbit pointer. Connecting with rabbitmqctl I get: | 21:41 |
| jms | * TCP connection succeeded but Erlang distribution failed | 21:41 |
| jms | * suggestion: hostname mismatch? | 21:41 |
| jms | * suggestion: is the cookie set correctly? | 21:41 |
| jms | The erlang packages are the same between hosts. | 21:41 |
| *** fthiagogv has quit IRC | 21:43 | |
| *** thangp has quit IRC | 21:43 | |
| *** annashen has joined #openstack-cinder | 21:48 | |
| *** bkopilov has joined #openstack-cinder | 21:49 | |
| *** mriedem has quit IRC | 21:51 | |
| *** daneyon has joined #openstack-cinder | 21:51 | |
| *** annashen has quit IRC | 21:53 | |
| *** dannywilson has quit IRC | 21:54 | |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Updating cmd/manage.py get_arg_string() argument parser https://review.openstack.org/186094 | 21:54 |
| *** dannywilson has joined #openstack-cinder | 21:55 | |
| uberjay | thingee: ping -- could I get some attention on our driver commit? I *think* everything is in order, just awaiting bp approval and core review. (https://review.openstack.org/#/c/178295/) | 21:58 |
| hemna | uberjay, is CI reporting ? | 21:58 |
| *** jwcroppe has quit IRC | 21:58 | |
| hemna | new driver? | 21:59 |
| uberjay | yes | 21:59 |
| uberjay | yes to both | 21:59 |
| *** jwcroppe has joined #openstack-cinder | 21:59 | |
| hemna | hey look at that. CI :) | 21:59 |
| hemna | uberjay, nice | 21:59 |
| uberjay | CI seems to be working. I've investigated a handful of failures, and I'm pretty sure they're legit | 21:59 |
| uberjay | hemna, thanks :) | 22:00 |
| hemna | uberjay, ok I have a meeting starting in like 1 minute ago. I'll see if I can look at it after. | 22:00 |
| uberjay | the CI is stable for the most part. every once in a while (once per week, or so), something seems to get jammed up in zuul or zuul-merger which causes the gerrit changes to stop being processed. | 22:02 |
| uberjay | hemna: awesome, thanks! i really appreciate it :) | 22:02 |
| *** akshai has joined #openstack-cinder | 22:03 | |
| *** mriedem has joined #openstack-cinder | 22:03 | |
| *** annashen has joined #openstack-cinder | 22:04 | |
| asselin | uberjay, a few tips: | 22:04 |
| asselin | 1. export GIT_BASE=https://review.openstack.org/p --> export GIT_BASE="https://git.openstack.org" | 22:05 |
| asselin | that other link will provide better performance | 22:06 |
| uberjay | ahh, ok. that may help; one of the failures looked to be a connectivity or performance issue while fetching a change | 22:07 |
| asselin | uberjay, well...just one. good job! :) | 22:08 |
| uberjay | how are they related to each other? can just treat them as equivalent? | 22:08 |
| uberjay | well, i suspect it'll get busier as June passes ;) | 22:09 |
| asselin | uberjay, yes, basically, the git.openstack.org link is a git farm that is load balanced | 22:09 |
| uberjay | ah, got it | 22:09 |
| asselin | review.openstack.org is gerrit which works, but not the best choice for CI | 22:09 |
| *** Rockyg has quit IRC | 22:11 | |
| uberjay | if you have any other tips, I'm all ears. :) | 22:11 |
| asselin | which repo are you using to setup your ci? | 22:12 |
| uberjay | it's all based on your os-ext-testing repo | 22:13 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/186112 | 22:15 |
| *** lpetrut has quit IRC | 22:19 | |
| *** annegentle has quit IRC | 22:20 | |
| asselin | uberjay, ok cool...glad to know that's working | 22:21 |
| *** bandwidth has quit IRC | 22:24 | |
| *** annegentle has joined #openstack-cinder | 22:24 | |
| *** crose has quit IRC | 22:28 | |
| anteaya | uberjay: thank you for starting out in the ci-sandbox repo | 22:28 |
| *** ndipanov has quit IRC | 22:30 | |
| *** tobe has joined #openstack-cinder | 22:30 | |
| uberjay | anteaya: no problem! considering the number of moving parts, it was clearly inappropriate to spam everyone with my setup process. ;) | 22:32 |
| anteaya | uberjay: yay | 22:32 |
| anteaya | someone with common sense | 22:32 |
| *** dannywil_ has joined #openstack-cinder | 22:32 | |
| anteaya | welcome!! | 22:32 |
| uberjay | haha ;) | 22:33 |
| uberjay | thanks :) | 22:33 |
| anteaya | thank you | 22:33 |
| *** kmartin_ has joined #openstack-cinder | 22:34 | |
| *** kmartin has quit IRC | 22:34 | |
| *** patrickeast_ has joined #openstack-cinder | 22:34 | |
| uberjay | asselin: yeah, it's working well. there were a couple moments when -infra diverged in incompatible ways while I was getting things going, but ... these things happen. I'm fairly certain without your help I'd still be scratching my head... | 22:34 |
| *** tobe has quit IRC | 22:35 | |
| *** dannywilson has quit IRC | 22:35 | |
| *** patrickeast has quit IRC | 22:36 | |
| *** patrickeast_ is now known as patrickeast | 22:36 | |
| asselin | uberjay, yeah...we're working on converging. if you'd like to help out, let me know! | 22:36 |
| uberjay | converging is an awesome and worthwhile goal. i'll absolutely let you know if/when I have the time. unfortunately, I just can't right now. | 22:38 |
| asselin | uberjay, no problem at all. | 22:38 |
| *** pboldin has quit IRC | 22:39 | |
| *** annashen has quit IRC | 22:43 | |
| *** annegentle has quit IRC | 22:46 | |
| *** jms has quit IRC | 22:46 | |
| *** annashen has joined #openstack-cinder | 22:50 | |
| *** krtaylor has quit IRC | 22:50 | |
| *** afazekas has quit IRC | 22:51 | |
| *** merooney has joined #openstack-cinder | 22:52 | |
| *** akshai has quit IRC | 22:52 | |
| nikeshm | uberjay: hi,are using trusty images in nodepool | 22:55 |
| openstackgerrit | Mitsuhiro Tanino proposed openstack/cinder-specs: Efficient volume copy for volume migration https://review.openstack.org/186209 | 22:55 |
| *** dannywil_ has quit IRC | 22:57 | |
| *** merooney has quit IRC | 22:59 | |
| openstackgerrit | Mitsuhiro Tanino proposed openstack/cinder-specs: Efficient volume copy for volume migration https://review.openstack.org/186209 | 22:59 |
| nikeshm | uberjay: what flavor are yu using for nodepool vms | 23:00 |
| *** angela-s has quit IRC | 23:00 | |
| uberjay | nikeshm: i am using trusty-based images (the cloud image) | 23:00 |
| nikeshm | uberjay: and flavor | 23:01 |
| *** Rockyg has joined #openstack-cinder | 23:01 | |
| uberjay | i created a new flavor between medium and large for the slave nodes... 6GB of ram & 2 VCPUs | 23:03 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 23:03 |
| *** mriedem has quit IRC | 23:04 | |
| uberjay | nikeshm: i needed a bit more -- 4GB is enough for devstack alone, but i'm running a self-contained installation of our storage software in a container on the same node. | 23:04 |
| *** manishg has left #openstack-cinder | 23:05 | |
| *** chlong has joined #openstack-cinder | 23:05 | |
| nikeshm | uberjay: ok actually i tried with 8 GB and 16GB RAM flavor but in nodepool vm its showing 6GB | 23:05 |
| nikeshm | uberjay: how much time its taking tto run tempest for yu | 23:06 |
| uberjay | you mean... if you ssh to one of the nodes it appears to have less memory? (where does it appear to have 6GB?) | 23:07 |
| nikeshm | yes | 23:08 |
| nikeshm | free -g | 23:08 |
| nikeshm | i tried after log in to nodepool node | 23:08 |
| nikeshm | its showing 6GB | 23:08 |
| uberjay | nikeshm: just the tempest portion of the test is "Ran: 281 tests in 711.0000 sec." | 23:08 |
| nikeshm | uberjay: cool!! ,what about devstack installation | 23:10 |
| uberjay | the rest of setup, etc, takes quite a bit of time though. in total, somewhere between 32-40 minutes for the full jenkins job | 23:10 |
| *** annashen has quit IRC | 23:12 | |
| uberjay | my slave nodes have ~6000MB of RAM, which is what i expect | 23:12 |
| nikeshm | uberjay: great,for me devstack itself taking 30 minutes | 23:13 |
| *** annashen has joined #openstack-cinder | 23:13 | |
| nikeshm | uberjay: how did yu cross-check that your system is using kvm | 23:13 |
| uberjay | nikeshm: huh... for me, the steps leading up to running devstack take ~4-5 minutes. running devstack, another 10 minutes or so, then tempest starts | 23:14 |
| *** ganso_ has quit IRC | 23:14 | |
| *** annashen has quit IRC | 23:17 | |
| uberjay | nikeshm: we do use kvm, do you think you might not be? | 23:21 |
| *** annashen has joined #openstack-cinder | 23:21 | |
| uberjay | that could certainly account for some performance issues... | 23:21 |
| nikeshm | uberjay: i am also using kvm,but in dashboard its showing qemu | 23:22 |
| nikeshm | uberjay: in yur dashboard of openstack did yu cross-check hypervisor details | 23:22 |
| *** sgotliv has quit IRC | 23:23 | |
| uberjay | ah, yeah, it does show qemu in the hypervisor summary. qemu makes use of the kvm feature, though. you should be able to check directly on the host if kvm is being used. i forget if it's visible through horizon? | 23:24 |
| nikeshm | uberjay: how can we check in host | 23:26 |
| *** annashen has quit IRC | 23:29 | |
| *** rmesta has quit IRC | 23:29 | |
| *** annashen has joined #openstack-cinder | 23:30 | |
| *** tobe has joined #openstack-cinder | 23:31 | |
| nikeshm | uberjay: did yu get any trouble with these tests http://paste.openstack.org/show/240850/ | 23:33 |
| *** asselin_ has joined #openstack-cinder | 23:34 | |
| uberjay | ah, you should be able to see a bunch of qemu-kvm processes running (one per guest). | 23:34 |
| nikeshm | asselin : able to run iscsi tests on my driver but some tests are giving trouble http://paste.openstack.org/show/240850/ | 23:34 |
| uberjay | you should see -machine accel=kvm, i guess you should ensure your cpu has virtualization support, but it would surprise me if it didn't! | 23:35 |
| *** annashen has quit IRC | 23:35 | |
| nikeshm | uberjay: qemu-system-x86_64 -enable-kvm -name instance-0000000d -S -machine pc-i440fx-trusty,accel=kvm | 23:36 |
| *** tobe has quit IRC | 23:36 | |
| asselin_ | nikeshm, I skip this one....don't know how to make it pass: tempest.thirdparty.boto.test_ec2_instance_run.InstanceRunTest.test_compute_with_volumes | 23:36 |
| asselin_ | nikeshm, this one I got working my using neutron networking instead of nova networking: tempest.scenario.test_minimum_basic.TestMinimumBasicScenario.test_minimum_basic_scenario | 23:37 |
| asselin_ | but it really depends on what the failure cause is...your mileage may vary | 23:37 |
| asselin_ | this one....I dont remember any issues...tempest.scenario.test_volume_boot_pattern.TestVolumeBootPattern.test_volume_boot_pattern although it fails intermittently | 23:38 |
| *** dims_ has quit IRC | 23:38 | |
| *** krtaylor has joined #openstack-cinder | 23:39 | |
| *** dims_ has joined #openstack-cinder | 23:40 | |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Updating cmd/manage.py get_arg_string() argument parser https://review.openstack.org/186094 | 23:40 |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Unit test for cmd/manage.py get_arg_string function https://review.openstack.org/186227 | 23:40 |
| uberjay | nikeshm: i have to run -- good luck! | 23:41 |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Unit test for cmd/manage.py get_arg_string function https://review.openstack.org/186227 | 23:43 |
| openstackgerrit | Daniel Allegood proposed openstack/cinder: Updating cmd/manage.py get_arg_string() argument parser https://review.openstack.org/186094 | 23:43 |
| *** vokt has quit IRC | 23:43 | |
| *** vokt has joined #openstack-cinder | 23:44 | |
| thingee | uberjay: read this https://wiki.openstack.org/wiki/Cinder/how-to-contribute-a-driver#Submitting_Driver_For_Review | 23:50 |
| *** leeantho has quit IRC | 23:50 | |
| *** Yogi1 has joined #openstack-cinder | 23:58 | |
| *** Yogi11 has joined #openstack-cinder | 23:58 | |
| *** julim has joined #openstack-cinder | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!