| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-07-18 | |||
| 14:41:44 | mriedem | pooja_jadhav: this specifically https://github.com/openstack/nova/blob/eb4f65a7951e921b1cd8d05713e144e72f2f254f/nova/compute/resource_tracker.py#L1455 | |
| 14:41:47 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Update queued-for-delete from the ComputeAPI during deletion/restoration https://review.openstack.org/566813 | |
| 14:42:01 | efried | mriedem: Not sure about "separate from dansmith's change". Gimme a minute here. | |
| 14:42:20 | mriedem | efried: dan's change is really about allocations in placement, | |
| 14:42:33 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: [POC] Graceful handling of nova-list when a cell is down https://review.openstack.org/567785 | |
| 14:42:36 | mriedem | there are still known issues with reporting disk usage, like if you have the DiskFilter enabled it's going to do a disk claim in the RT | |
| 14:42:48 | mriedem | we can't drop the disk claims until we drop caching scheduler | |
| 14:42:54 | mriedem | since it doesn't use placement | |
| 14:43:10 | mriedem | fixing https://github.com/openstack/nova/blob/eb4f65a7951e921b1cd8d05713e144e72f2f254f/nova/compute/resource_tracker.py#L1455 is going to be non-trivial i think | |
| 14:44:00 | mriedem | well maybe it's not that bad, we have an instance uuid in scope when it's used and from the instance we can determine if it's volume-backed or not | |
| 14:47:10 | efried | mriedem: How do I access the compute node object from the compute service thingy (self.compute1)? | |
| 14:49:13 | pooja_jadhav | mriedem: you mean to say 'root_gb' is taking directly from flavor.root_gb. whether we create instance from volume or not? correct? | |
| 14:49:17 | mriedem | efried: in a functional test? | |
| 14:49:23 | mriedem | pooja_jadhav: correct | |
| 14:49:31 | efried | mriedem: yeah | |
| 14:49:36 | mriedem | efried: self.compute1 is the service record, | |
| 14:50:36 | mriedem | from that you can get the manager and host i think, and from that you can get the compute node record by hostname and then get it's uuid | |
| 14:50:42 | mriedem | self.compute1.manager.host | |
| 14:51:15 | mriedem | that's how ComputeManager._get_compute_nodes_in_db works anyway | |
| 14:51:16 | efried | mriedem: That's just a string. | |
| 14:52:04 | mriedem | right to lookup the compute node | |
| 14:52:14 | pooja_jadhav | mriedem: in that case, we need to check, if the instance is volume_backed then make root_gb is to 0. | |
| 14:52:24 | mriedem | i guess you could use self.compute1.manager._resource_tracker.compute_nodes | |
| 14:52:30 | mriedem | assuming those are set by the time you access the thing | |
| 14:52:37 | mriedem | pooja_jadhav: yes i understand the issue, | |
| 14:52:45 | mriedem | pooja_jadhav: but i don't think it needs to be done in dansmith's change | |
| 14:52:57 | mriedem | it's a separate reporting issue which can be fixed on top of dan's change | |
| 14:53:18 | pooja_jadhav | mriedem: Ohk | |
| 14:53:28 | efried | mriedem: it is, thanks. | |
| 14:53:40 | mriedem | pooja_jadhav: but if you can report the nova bug that would be helpful | |
| 14:53:48 | efried | pooja_jadhav: Let me know when you have that bug open. I'm working on the test case for it. | |
| 14:54:29 | pooja_jadhav | efried: Ok sure. | |
| 14:54:31 | mriedem | i'm going to rebase dan's change | |
| 14:54:32 | sahid | jaypipes: did you had a chance to look at this os_acc project? | |
| 14:54:35 | sahid | https://review.openstack.org/#/c/583097/ | |
| 14:54:54 | efried | mriedem: ack, fwiw there were no conflicts when I did that locally, even though gerrit complained. | |
| 14:54:56 | mriedem | i've been meaning to follow up on dan's change and fix the move operation stuff too | |
| 14:55:10 | sahid | it seems that they want to build the guest configuration (for libvirt domain XML) | |
| 14:55:13 | efried | mriedem: mebbe there's something I need to pull on master since last night. | |
| 14:55:27 | efried | mriedem: If it does wind up being manual, you wanna fix those nits I noted? | |
| 14:55:40 | mriedem | efried: yeah i'll fix those | |
| 14:55:48 | jaypipes | sahid: not really yet, sorry :( | |
| 14:56:10 | sahid | jaypipes: no no worries, it's just that i don't want give to them bad direction | |
| 14:56:24 | mriedem | yeah stephenfin's RequestSpec.network_metadata change landed | |
| 14:56:39 | kashyap | sahid: What is that project even? | |
| 14:57:55 | sahid | it's not really clear for me, my thinking is Cyborg would be that generic device manager we are missing and os-acc the data model used to talk between cyborg and nova | |
| 14:58:02 | jangutter | kashyap: My personal lossy summary of Cyborg: How to cloud-ify FPGA's. | |
| 14:59:11 | jangutter | kashyap, sahid: the first use-case is for FPGA's, and hopefully it should be a general device manager. The idea is that os-acc should function as the object model and a plugin model for Nova. | |
| 14:59:40 | kashyap | jangutter: "Someone" needs to write-up a damn good spec with a clear beginning, middle and an end. | |
| 15:00:00 | sahid | jangutter: ok thanks for i have well understood the thing :) | |
| 15:00:04 | efried | kashyap: There are specs. | |
| 15:00:06 | kashyap | jangutter: But noted :-) | |
| 15:00:25 | efried | kashyap: They're in the cyborg repo, not the cyborg-specs repository, confusingly. | |
| 15:00:28 | kashyap | efried: Oh, I'll have to look it up | |
| 15:00:42 | jangutter | kashyap: I accidentally got this info via osmosis in Dublin. | |
| 15:00:49 | kashyap | Also that patch should be broken down and made into series with great (yes, great) commit messages. | |
| 15:01:40 | jangutter | kashyap: hehe, first review comment should be a link to the spec, probably. | |
| 15:01:40 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Add gabbi coverage for an inventory change https://review.openstack.org/583642 | |
| 15:01:41 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Add gabbi coverage for inv of missing rp https://review.openstack.org/583643 | |
| 15:01:42 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] cover bad content-length header https://review.openstack.org/583644 | |
| 15:02:23 | efried | kashyap: doc/source/specs in openstack/cyborg, and one open patch: https://review.openstack.org/#/c/577438/ | |
| 15:07:01 | mriedem | efried: can i deal with the changed link in https://review.openstack.org/#/c/515423/29 in a follow up? also your concern about the test isn't something i'm worried about - the dicts in question have only a single key | |
| 15:10:21 | efried | mriedem: okay, +2. | |
| 15:10:39 | mriedem | thanks; looks like there are other things in the series from which i can draw a fup cleanup patch | |
| 15:12:21 | efried | mriedem: Are you in the middle of that rebase or can I push that test case on top and have you carry it along? | |
| 15:12:38 | efried | pooja_jadhav: How's that bug report coming? | |
| 15:13:34 | mriedem | efried: pushing now | |
| 15:14:15 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Avoid requesting DISK_GB allocation for root_gb on BFV instances https://review.openstack.org/580720 | |
| 15:18:14 | openstackgerrit | Eric Fried proposed openstack/nova master: WIP: Func test for improper local DISK_GB reporting https://review.openstack.org/583646 | |
| 15:18:24 | efried | mriedem, pooja_jadhav: ^ | |
| 15:18:29 | efried | Just need to fill in that bug number. | |
| 15:20:59 | mriedem | efried: funny it doesn't report swap disk usage | |
| 15:23:27 | mriedem | comments inline, but lgtm | |
| 15:28:39 | mriedem | so it appears that going back to grizzly when the DiskFilter was added, we only accounted for local disk usage using root_gb and ephemeral_gb, not swap https://review.openstack.org/#/c/13182/ | |
| 15:30:49 | mriedem | the DiskFilter was fixed in Icehouse but the RT wasn't https://review.openstack.org/#/c/51323/ | |
| 15:32:08 | efried | mriedem: That /os-hypervisors/statistics API isn't working, something missing in the fixture setup. Worth fixing that? | |
| 15:33:02 | mriedem | efried: not working how? | |
| 15:33:08 | mriedem | are you using the admin_api client? | |
| 15:33:47 | efried | mriedem: http://paste.openstack.org/show/726208/ | |
| 15:33:53 | efried | Probably not :) | |
| 15:35:12 | efried | mriedem: Same thing via the admin_api | |
| 15:35:22 | efried | Looking at the func test that actually does this... | |
| 15:35:57 | mriedem | /home/efried/Neo/nova/.tox/functional | |
| 15:35:57 | mriedem | neo huh | |
| 15:36:00 | mriedem | tsk tsk | |
| 15:36:22 | efried | mriedem: swhat our project used to be called before it was called NovaLink | |
| 15:36:59 | mriedem | oh i know | |
| 15:37:19 | mriedem | i knew about neo before it couldn't be talked about openly | |
| 15:37:25 | mriedem | i'm a neo hipster really | |
| 15:39:06 | efried | 'cept you can't grow a beard | |
| 15:39:09 | mriedem | i don't see anything special in the os-hypervisors api samples functional tests for the statistics call | |
| 15:39:19 | mriedem | hipster bears are on their way out | |
| 15:39:23 | mriedem | *beards | |
| 15:39:27 | efried | mriedem: Me neither. I thought it might be the json headers, but I tried adding those and it didn't help. | |
| 15:39:44 | efried | So what you're saying is you were a beardless hipster before it was cool. | |
| 15:40:43 | efried | ya know, it's possible I'm injecting unintended side effects by using remote_pdb. Lemme try unwinding that... | |
| 15:41:11 | mriedem | efried: you just blew my mind | |
| 15:43:24 | melwitt | mriedem: we have a bug where the libvirt driver will unconditionally destroy disks even on shared storage during a rebuild failure, and we're looking at how to write the fix in such a way that it's backportable. what we really need to know in the libvirt driver spawn() is whether recreate=True or destroy_disks_on_failure should be True/False. would such a change to the spawn method signature be banned for a backport? | |
| 15:43:28 | efried | mriedem: That was it, though I don't know why. | |
| 15:44:15 | melwitt | mriedem: this is the WIP of the change, avoiding the method signature change, and it duplicates code from compute manager's default rebuild impl into the libvirt driver https://review.openstack.org/578846 | |