| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-02-20 | |||
| 10:34:49 | bauzas | mdbooth: others, I need to disappear, but I'm tempted to set https://bugs.launchpad.net/nova/+bug/1740025 as Wontfix | |
| 10:34:50 | openstack | Launchpad bug 1740025 in OpenStack Compute (nova) "LVM Sparse Volumes broken" [Undecided,New] | |
| 10:34:51 | bauzas | agreed ? | |
| 10:35:34 | mdbooth | bauzas: Reading | |
| 10:36:22 | mdbooth | bauzas: Why wontfix? | |
| 10:36:55 | mdbooth | I recall there being severe problems with sparse lvm | |
| 10:37:03 | mdbooth | Not the specifics, though | |
| 10:38:53 | bauzas | mdbooth: because it looks like a feature | |
| 10:39:21 | bauzas | mdbooth: tbc, you can do thin provisioning with Cinder | |
| 10:39:23 | kashyap | I don't know, would you want the reporter to do the whole spec dance? | |
| 10:39:36 | bauzas | but here, the reporter only uses raw LVM | |
| 10:39:37 | kashyap | It's okay to flesh out the details on the bug, without hurrying to close it | |
| 10:39:47 | mdbooth | bauzas: It reads like a bug to me. | |
| 10:39:51 | kashyap | Yeah, to me too. | |
| 10:40:05 | bauzas | so, in case you create LVM volumes having X size | |
| 10:40:08 | bauzas | then, meh | |
| 10:40:15 | mdbooth | Sparse LVM is a thing, and if we were using it correctly, I would expect the scenario described to work. | |
| 10:40:39 | bauzas | mdbooth: why can't the reporter create LVs that are more than 64MB ? | |
| 10:41:01 | bauzas | that's something I don't understand | |
| 10:41:02 | mdbooth | bauzas: I'm guessing because we didn't create the sparse lv correctly. | |
| 10:41:20 | lyarwood | yeah that's not up to the user, n-cpu is creating the lv's | |
| 10:41:20 | mdbooth | bauzas: As I said, though, I'm pretty sure I recall severe problems in our implementation of sparse lvm. | |
| 10:41:29 | bauzas | ah ok | |
| 10:41:34 | mdbooth | severe == How could it ever possibly have worked | |
| 10:41:41 | bauzas | so when nova creates the lv, it asks for a size | |
| 10:41:48 | bauzas | meh | |
| 10:41:48 | bauzas | that's not the image | |
| 10:42:04 | mdbooth | bauzas: It's like creating a sparse file, except lvm | |
| 10:42:12 | bauzas | and then tries to thin provision by itself, while cinder can do that far better | |
| 10:42:31 | mdbooth | It creates a sparse block device, which only allocated blocks from the lvm block pool when they're used | |
| 10:42:35 | mdbooth | Or it's supposed to, anyway | |
| 10:42:36 | bauzas | looks to me we barely support it | |
| 10:42:44 | mdbooth | bauzas: Right. | |
| 10:42:54 | bauzas | yeah I understand the purpose of sparse LVs | |
| 10:42:59 | mdbooth | bauzas: I'm not convinced this has ever worked. However, I'm pretty sure that it was intended to work. | |
| 10:43:05 | bauzas | exactly | |
| 10:43:08 | bauzas | hence the Wontfix | |
| 10:43:14 | bauzas | we never tested that, right? | |
| 10:43:25 | mdbooth | Meh, still sounds like a bug to me. | |
| 10:43:29 | bauzas | some obscure blueprint that implemented in 2012 to make people happy | |
| 10:43:37 | bauzas | without having decent testing | |
| 10:43:48 | bauzas | call it tech debt | |
| 10:43:49 | mdbooth | If we mark this wontfix, we should officially deprecate the feature instead. | |
| 10:43:54 | bauzas | that's right | |
| 10:44:08 | bauzas | I need to go to the gym, but I'll consider all the options before replying to the bug | |
| 10:44:21 | mdbooth | bauzas: KK. Enjoy! | |
| 10:44:29 | bauzas | mdbooth: yeah, we can have volume-backed instances by cinder, right? | |
| 10:44:37 | mdbooth | bauzas: And don't forget I'll be Wattbiking in Dublin next week if you want to join me :) | |
| 10:44:41 | bauzas | mdbooth: if so, you can do thin provisioning with cinder | |
| 10:44:48 | kashyap | mdbooth: What _is_ Wattbiking? /me looks up | |
| 10:45:00 | lyarwood | bauzas: yeah but not using local disks as PVs as this is doing | |
| 10:45:03 | kashyap | Ah, indoor bike trainers | |
| 10:45:06 | mdbooth | bauzas: Yes, I'm sure this is much better handled by cinder. | |
| 10:45:23 | bauzas | mdbooth: this is what I'm doing https://vimeo.com/8178409 | |
| 10:47:29 | kashyap | Bail out whom? | |
| 11:04:34 | mdbooth | kashyap: See the second definition of bail out in that list | |
| 11:04:43 | kashyap | I saw that... | |
| 11:04:47 | kashyap | The parachute thing, right? | |
| 11:04:59 | kashyap | No | |
| 11:05:02 | kashyap | "to abandon a harmful or difficult situation;" | |
| 11:05:03 | kashyap | :P | |
| 11:05:10 | kashyap | mdbooth: I doubt bauzas mean't that. Haha | |
| 11:05:32 | kashyap | (I'd guess, he'd be surprised to learn that.) | |
| 11:05:50 | mdbooth | Not used literally, it seems appropriate in context | |
| 11:05:51 | kashyap | s/mean't/meant/ | |
| 11:06:47 | kashyap | Yeah, true. | |
| 11:33:08 | openstackgerrit | Marcin Juszkiewicz proposed openstack/nova master: Allow to configure amount of PCIe ports https://review.openstack.org/545034 | |
| 13:30:08 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: RP delete inventories (v1.5) https://review.openstack.org/514642 | |
| 13:30:08 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: RP list: member_of and resources parameters (v1.3, v1.4) https://review.openstack.org/511183 | |
| 13:30:09 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: Resource class set (v1.7) https://review.openstack.org/514644 | |
| 13:30:09 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: CLI for traits (v1.6) https://review.openstack.org/514643 | |
| 13:30:10 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: CLI allocation candidates (v1.10) https://review.openstack.org/514647 | |
| 13:30:10 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: Usages per project and user (v1.8, v1.9) https://review.openstack.org/514646 | |
| 13:30:11 | openstackgerrit | Bence Romsics proposed openstack/osc-placement master: New dict format of allocations (v1.11, v1.12) https://review.openstack.org/542819 | |
| 13:46:02 | openstackgerrit | Elod Illes proposed openstack/nova master: Functional test: cold migrate to compute down https://review.openstack.org/496280 | |
| 13:46:21 | openstackgerrit | Elod Illes proposed openstack/nova master: Functional test: evacuate with no compute https://review.openstack.org/498482 | |
| 14:10:51 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/queens: Add the ability to get absolute limits from Cinder https://review.openstack.org/546157 | |
| 14:10:52 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/queens: Check quota before creating volume snapshots https://review.openstack.org/546158 | |
| 14:12:08 | mriedem | claudiub: can you get these backports? https://review.openstack.org/#/q/topic:bug/1746032+status:open | |
| 14:44:49 | jaypipes | is it Friday yet? | |
| 14:45:40 | dansmith | no | |
| 14:45:57 | dansmith | I can tell because I'm not in a big metal tube with sick people and crying children | |
| 14:46:27 | efried | ^ | |
| 14:46:47 | efried | Anybody got advice for what a guy can do with a 7h layover in NYC? | |
| 14:47:09 | efried | Noting that both the guy in question and his spouse pretty much loathe NYC | |
| 14:47:25 | jaypipes | efried: which airport? | |
| 14:48:57 | efried | jaypipes: NYC | |
| 14:49:17 | efried | is that an airport? | |
| 14:51:27 | jroll | likely JFK | |
| 14:52:03 | mriedem | you can forget about it | |
| 14:52:12 | efried | Oh, it is JFK | |
| 14:52:14 | mriedem | need another core on https://review.openstack.org/#/c/340614/ - it's not as scary as it looks | |
| 14:52:22 | mriedem | ^ has to go back to ocata before eol | |
| 14:52:25 | efried | I thought NYC wasn't an airport. Knew about JFK and LGA... | |
| 14:53:11 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Convert driver supported capabilities to compute node provider traits https://review.openstack.org/538498 | |
| 14:53:34 | mriedem | my flight to amsterdam is half empty | |
| 14:54:07 | efried | Some might say half full. | |
| 14:54:19 | efried | Interesting inversion of the usual optimist/pessimist viewpoints, that. | |
| 14:54:35 | mriedem | in the case of a flight, i want it to be half empty | |
| 14:54:45 | efried | exactly | |