| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-05-23 | |||
| 20:05:33 | penick | Ugh ceph. | |
| 20:05:54 | melwitt | yeah, you might be thinking of backing instance ephemeral storage on cinder (ceph?). boot from volume is creating a volume in cinder from an image and attaching it to an instance, and it (the volume) can live on after the instance is deleted | |
| 20:05:54 | mriedem | fine fine | |
| 20:05:56 | mriedem | use scaleio! | |
| 20:06:26 | mriedem | melwitt: or not, delete_on_termination ftw | |
| 20:06:28 | dansmith | 8GB swap for all! | |
| 20:06:32 | penick | So Nova would interact with something like scaleio directly, and not go through cinder? | |
| 20:06:39 | melwitt | mriedem: yep, can do that too | |
| 20:07:04 | dansmith | penick: this is the thing we were talking about -- nova duplicates a fraction of cinder for the drivers we like | |
| 20:07:23 | dansmith | penick: so you can back your instances on ceph, or local, but not much else | |
| 20:07:32 | dansmith | because we like writing things twice and maintaining them once | |
| 20:07:52 | cdent | can I quote you on that dansmith? that's very quotable | |
| 20:07:58 | penick | But..why not let Cinder do the word of managing volumes? | |
| 20:08:06 | dansmith | cdent: you may | |
| 20:08:07 | dansmith | penick: yeah, that | |
| 20:08:12 | dansmith | has been my suggestion, | |
| 20:08:15 | dansmith | but it's a ton of work | |
| 20:08:30 | dansmith | and openstack being already dead according to the press, we don't have enough time to finish that before we're done | |
| 20:08:46 | penick | Pfft. Dead. Hah | |
| 20:08:58 | dansmith | regardless, I think we all agree that that's how it should work | |
| 20:09:03 | dansmith | but it doesn't today | |
| 20:09:12 | dansmith | unless you BFV everything, which is also broken for lots of things | |
| 20:09:20 | dansmith | and, nova devs will stop talking to you if you move to that | |
| 20:10:09 | penick | So if I add a feature my users need, in the only way that it works right now, you'll kick me out of nova hogworts? | |
| 20:10:19 | mriedem | penick: i'm much more interested in your doing cells v2 from trunk than doing BFV | |
| 20:10:34 | dansmith | penick: don't worry, you'll hate yourself soon after adopting it anyway | |
| 20:11:55 | penick | mriedem: i'm much more interested in cellsv2 as well. But to quote a great poet "If you don't eat your meat, you can't have any pudding" | |
| 20:12:06 | penick | Not sure who said that. Might have been Socrates. | |
| 20:12:14 | dansmith | pretty sure it was Waters :P | |
| 20:12:23 | mriedem | dude, | |
| 20:12:26 | penick | :P | |
| 20:12:28 | mriedem | you don't want to eat boiled meat anyway right? | |
| 20:12:51 | mriedem | where is mdbooth when i need to joke about terrible british meat | |
| 20:13:35 | penick | Correct, but if I picked the things I -wanted- to do, over what the business needs, i'd have a really kickass 10 node openstack cluster running on trunk. But I wouldn't be moving the whole company over to it | |
| 20:23:48 | mriedem | well riddle me this, where is accessIPv4 actually used anywhere in nova anymore?! | |
| 20:24:51 | mriedem | i need an onboarding session | |
| 20:24:57 | cburgess | mriedem Thats a dangerous thing to say to me. :P | |
| 20:25:09 | mriedem | cburgess: whatever i said, it was probably a day ago, | |
| 20:25:10 | mriedem | so remind me | |
| 20:25:19 | cburgess | mriedem Getting enough cloud providers to ask for volume type. | |
| 20:25:26 | mriedem | oh right, page 45 | |
| 20:25:34 | cburgess | page 45? | |
| 20:25:58 | mriedem | page 45 of the nova core dungeon master's guide states: "once you get enough people asking for the same feature, regardless of policy, it must be accepted" | |
| 20:26:06 | cburgess | lol | |
| 20:26:09 | penick | Pretty sure you just made that up | |
| 20:26:15 | cburgess | I didn't know such a guide exists. Must be for PTLs only. | |
| 20:26:23 | mriedem | "nova core dungeon master's guide" | |
| 20:26:34 | cburgess | Oh core | |
| 20:26:35 | cburgess | My bad.. | |
| 20:26:43 | mriedem | penick: melwitt's is keeping your dining table steady atm | |
| 20:26:45 | dansmith | I haven't seen this guide | |
| 20:26:49 | cburgess | mriedem You do realize that has to become a think now right? | |
| 20:27:01 | penick | We have a dining table? | |
| 20:27:04 | cburgess | melwitt is doing what now? | |
| 20:27:11 | dansmith | isn't that melwitt's workstation now? | |
| 20:27:15 | melwitt | what is happening | |
| 20:27:28 | mriedem | i know what's not happening | |
| 20:27:43 | mriedem | ... https://review.openstack.org/#/c/446239/ | |
| 20:27:44 | mriedem | :) | |
| 20:28:00 | melwitt | it's happening, at the subaru dealership even! | |
| 20:28:05 | mriedem | oh snap | |
| 20:28:18 | melwitt | about to upload the quotas bomb again | |
| 20:28:25 | penick | her workstation locality depends on three factors. The season, is the cat being annoying, or is she at the dealership | |
| 20:28:28 | mriedem | page 62 has a subaru coupon coincidentally | |
| 20:28:31 | mriedem | on free granola bar | |
| 20:28:33 | mriedem | *one | |
| 20:28:45 | penick | free granola bar, that's a sweet deal | |
| 20:28:53 | melwitt | I ate a free Cliff bar here earlier | |
| 20:28:58 | mriedem | haha | |
| 20:29:26 | mriedem | i always feel great when some joke i make which might be offensive turns out to be actual fact | |
| 20:29:36 | melwitt | lol, right? | |
| 20:29:48 | penick | man that nova dungeon masters guide is handy | |
| 20:29:57 | mriedem | why do you think we don't add more cores? | |
| 20:30:01 | mriedem | limited supply | |
| 20:30:03 | mriedem | of guides | |
| 20:32:02 | penick | Not cheap to print. also the "Worlds best nova core" mugs don't exactly grow on trees. Those legbreakers at Cafepress charge a fortune | |
| 20:32:12 | melwitt | merge conflicts in local delete code ... grrrr | |
| 20:32:24 | mriedem | melwitt: blame dansmith | |
| 20:32:53 | melwitt | >:| angry eyebrows | |
| 20:33:18 | dansmith | go ahead | |
| 20:33:26 | dansmith | I'll flip this table and leave | |
| 20:34:13 | melwitt | the workstation table might be hard to flip. but could be toppled over | |
| 20:34:49 | penick | Cool cores don't watch tables flip. They just walk away. (-_- )ﾉ⌒┫ ┻ ┣ | |
| 20:35:18 | dansmith | well, I'm not going to flip _this_ table since it has all my monitors and laptop on, but I will go find a suitable table with minimal collateral damage and flip that mofo | |
| 20:35:35 | melwitt | :D | |
| 20:36:07 | melwitt | we're all on the same page here | |
| 20:36:40 | cburgess | dansmith pics or it didn't happen | |
| 20:36:45 | mriedem | you know what i think we can all agree on? | |
| 20:36:51 | mriedem | removing the gluster volume driver https://review.openstack.org/#/c/463992/ | |
| 20:36:59 | cburgess | +1 LGTM | |
| 20:37:02 | dansmith | heh | |
| 20:39:52 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Cleanup update_instance cell mapping handling https://review.openstack.org/467380 | |
| 20:39:53 | mriedem | dansmith: i'm also hoping you don't find problems with ^ | |
| 20:40:10 | mriedem | i think that stuff all predates _get_instance storing the targeted cell context on the instance object | |
| 20:40:41 | dansmith | mriedem: can we wait and or depends-on that on the multicell devstack patch? | |
| 20:40:56 | mriedem | i can put a change on top with the depends-on to test it out | |
| 20:41:02 | dansmith | okay | |
| 20:41:19 | dansmith | or just dependson it and remove it if it's good | |
| 20:41:20 | mriedem | https://review.openstack.org/#/c/436094/ right? | |
| 20:41:28 | dansmith | yar | |