| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-08 | |||
| 21:08:14 | dansmith | cburgess: right | |
| 21:08:19 | melwitt | but, if the timeout is really 2 minutes then I could see the --graceful option being an option | |
| 21:08:45 | cburgess | OK well sorry to leave in the middle of this but wife just called to say the hot water heater is making a very loud banging noise continually.. so... yeah... thats fun. | |
| 21:09:08 | dansmith | melwitt: to me "delete" does not mean "don't really delete", but.. we lost the actually-deleting-things battle a while ago :P | |
| 21:09:36 | sdague | the main reason I think a lot of people want this is windows shutting down, and you get the "I'm applying patches don't turn me off" screen | |
| 21:09:51 | sdague | melwitt: so the timeout has to be at least 2 minutes so you don't kill that | |
| 21:09:58 | dansmith | oh forgot about that, that's a good one | |
| 21:10:00 | melwitt | dansmith: heh, I guess. as a layperson, I didn't think of "delete" as meaning much other than "get rid of this instance" and not thinking about what happens underneath | |
| 21:10:45 | mriedem | i think of delete as 'get rid of this instance' but also, don't mess up all my crap that's attached to it unless you're also deleting that, like volumes and ports | |
| 21:10:47 | cfriesen | isn't the default shutdown timeout 60 sec? | |
| 21:10:53 | melwitt | 2 minutes for windows update, that's generous (to think it would be that short) | |
| 21:11:06 | dansmith | melwitt: I know, 2 minutes is laughable | |
| 21:11:08 | sdague | melwitt: once upon a time there was 120 timer here | |
| 21:11:27 | dansmith | mriedem: "mess up my crap" is relative though. it won't corrupt the filesystem | |
| 21:11:30 | sdague | and it did do all sorts of oddity in the gate, because cirros totally ignores acpi | |
| 21:11:38 | sdague | so we were hanging around forever | |
| 21:11:49 | dansmith | mriedem: if your app corrupts your data on this event, then it also corrupts your data if the power goes out in the datacenter, so.. meh | |
| 21:11:52 | mriedem | i vaguely remember that, | |
| 21:12:00 | mriedem | didn't we revert Phil's like 1K LOC patch or somethign? | |
| 21:12:03 | melwitt | if it were called 'nova pull-the-power-cord' that would be clear to me :P | |
| 21:12:11 | mriedem | dansmith: true | |
| 21:12:22 | sdague | mriedem: maybe, I don't remember exactly how long it was around | |
| 21:12:32 | dansmith | sdague: ah, that's what I was thinking of | |
| 21:12:34 | melwitt | yeah, I can see both arguments | |
| 21:12:59 | mriedem | so at what point so volume multi-attach fix all of this? :) | |
| 21:13:05 | mriedem | s/so/does/ | |
| 21:13:35 | sdague | so, if it was a thing... (which I'm still not convinced it should be) it definitely doesn't seem like a system wide answer is the right answer | |
| 21:13:48 | dansmith | mriedem: volume multiattach asumes you don't have an app that would corrupt things like this, so that's a good argument for not caring IMHO :) | |
| 21:13:54 | sdague | it would be image based or instance based metadata for shutdown time | |
| 21:14:15 | sdague | so windows people could tune it up to 20 minutes for snowflakes | |
| 21:14:17 | cfriesen | sdague: you mean like os_shutdown_timeout ? | |
| 21:14:17 | dansmith | image.I_suck=True | |
| 21:14:33 | sdague | cfriesen: maybe, that already in there? | |
| 21:14:40 | cfriesen | sdague: yep | |
| 21:14:41 | sdague | that vaguely sounds familiar | |
| 21:14:43 | dansmith | lol | |
| 21:14:48 | sdague | ok, probablem solved! | |
| 21:14:58 | dansmith | unfuckingbelievable | |
| 21:15:04 | cfriesen | sdague: well, except it's only for shutdown, not for delete | |
| 21:15:04 | melwitt | hahaha | |
| 21:15:04 | dansmith | I want my 20 minutes back | |
| 21:15:47 | sdague | cfriesen: ok, but that means there is still a way for you to get what you want right? | |
| 21:15:48 | dansmith | ah, so setting it doesn't imply the different behavior on delete | |
| 21:15:56 | sdague | delete is delete | |
| 21:16:10 | cfriesen | dansmith: correct. sdague: I can manually do a nova shutdown before doing a delete, sure. | |
| 21:16:22 | mriedem | we don't check the image meta / extra spec for that on delete, only stop | |
| 21:16:26 | dansmith | right | |
| 21:16:35 | mriedem | but we could... | |
| 21:16:44 | sdague | ok, so I feel like thats fine then, there is an API people can use if they want the behavior | |
| 21:16:45 | mriedem | w/o new API parameters | |
| 21:16:57 | dansmith | sdague: agreed | |
| 21:17:17 | dansmith | sdague: we knew stop does what we want, but didn't know it was configurable timeout, which makes it even more useful for the thing | |
| 21:17:18 | mriedem | i always forget image_meta/extra_specs | |
| 21:17:31 | sdague | mriedem: well, they also aren't hugely tested | |
| 21:17:41 | mriedem | sdague: neither are the 5K config options | |
| 21:17:48 | dansmith | I think you mean bigly | |
| 21:17:50 | mriedem | but i tend to think of them in the same way | |
| 21:17:54 | cfriesen | sdague: maybe we document that "nova delete" is explicitly a hard shutdown, and may result in data loss unless the guest flushes disks first? | |
| 21:18:51 | openstackgerrit | Pushkar Umaranikar proposed openstack/nova master: Placement api: set custom json_error_formatter in allocations https://review.openstack.org/395194 | |
| 21:19:05 | sdague | http://developer.openstack.org/api-ref/compute/?expanded=stop-server-os-stop-action-detail,delete-server-detail#delete-server | |
| 21:19:12 | sdague | "By default, the instance is going to be (hard) deleted immediately from the system" | |
| 21:19:31 | sdague | it could be tightened up for sure | |
| 21:19:51 | dansmith | By default, the instance is going to be (hard) deleted immediately | |
| 21:19:52 | dansmith | yeah | |
| 21:19:55 | mriedem | oh nice | |
| 21:19:56 | mriedem | "With correct permissions, you can see the server status as" | |
| 21:19:56 | melwitt | yeah, I think it wouldn't hurt and only help to add that bit of additional obvious info | |
| 21:19:58 | mriedem | AS WHAT?! | |
| 21:20:17 | melwitt | it's a secret! | |
| 21:20:40 | sdague | mriedem: we have the best statuses | |
| 21:20:45 | mriedem | i paid for the whole seat but only need the edge | |
| 21:20:52 | dansmith | that's true, if you really want this a an op, you can turn on soft_delete for 2 minute reclaim | |
| 21:21:11 | sdague | dansmith: well... that's different yet again right? | |
| 21:21:16 | sdague | That just defers the delete | |
| 21:21:24 | sdague | it still hard deletes when it happens | |
| 21:21:25 | mriedem | does it stop the instance? | |
| 21:21:26 | dansmith | sdague: same effect I think, we shutdown the instance and then delete it two minutes later | |
| 21:21:34 | sdague | I thought we just held | |
| 21:21:37 | dansmith | I don't think it hard deltes at the beginng | |
| 21:21:38 | sdague | it was more for the undelete case | |
| 21:21:48 | dansmith | no, it doesn't keep running | |
| 21:21:49 | dansmith | sdague: it is | |
| 21:22:24 | dansmith | self.driver.power_off(instance) | |
| 21:22:42 | mriedem | yup[ | |
| 21:22:44 | dansmith | graceful shutdown, which will wait, and then soft_delete will reclaim in $interval | |
| 21:22:45 | mriedem | was just looking at that | |
| 21:23:15 | sdague | ah, ok | |
| 21:23:24 | mriedem | i was recently complaining about reclaim_instance_interval | |
| 21:23:46 | mriedem | but in my old age i've forgotten why | |
| 21:23:55 | dansmith | because it's stupid? | |
| 21:24:14 | mriedem | well, my complaint was if you set reclaim_instance_interval in the api but not on the compute hosting the instance, the instance won't actually be deleted | |
| 21:24:20 | mriedem | but, we assume you don't do that | |
| 21:24:24 | dansmith | ah yep | |
| 21:24:49 | mriedem | we assume you've correctly configured nova.conf's 5K options and deployed it correctly on all of your services | |
| 21:24:59 | mriedem | oops mine just blew off your toe! | |
| 21:25:38 | mriedem | at some point i think capabilities will bridge that gap | |
| 21:25:48 | mriedem | since the api could tell what the capabilities are on the host | |
| 21:25:51 | cfriesen | looks like only xen supports soft delete | |
| 21:26:17 | cfriesen | oh, I guess we fall back to just power_off | |
| 21:26:18 | dansmith | cfriesen: we fall back to poweroff if the driver doesn't do anything | |