| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-04-17 | |||
| 10:00:22 | alex_xu | cdent: should we return 204 for put on a standard resource class? | |
| 10:01:25 | openstackgerrit | Chris Dent proposed openstack/nova master: Update resource tracker to PUT custom resource classes https://review.openstack.org/456915 | |
| 10:01:26 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Idempotent PUT /resource_classes/{name} https://review.openstack.org/448791 | |
| 10:01:31 | alex_xu | cdent: I think no, it doesn't help for client code | |
| 10:02:51 | cdent | alex_xu: I think it is safe to not do it, but it does make for a bit of an asymmetry | |
| 10:03:21 | cdent | you can GET /resource_classes/VCPU but not PUT /resource_classes/VCPU | |
| 10:04:58 | cdent | If we were being super strict, then both should work | |
| 10:05:41 | alex_xu | cdent: the nova client code will know the standard rc, so it wastes to send another put request. so at least, nova won't send put on standard rc. | |
| 10:05:54 | openstackgerrit | shaofeng cheng proposed openstack/nova master: Update swift version newton to ocata https://review.openstack.org/457197 | |
| 10:07:07 | alex_xu | for the anyone use placement out of nova, they should hard-code those standard rc in code again? | |
| 10:07:17 | alex_xu | at least we only have nova use placement | |
| 10:07:20 | alex_xu | for now | |
| 10:08:09 | cdent | I think for now, it is okay, but we may need to make some adjustments in the future, but that's okay too | |
| 10:08:46 | alex_xu | cdent: yea | |
| 10:08:55 | openstackgerrit | shaofeng cheng proposed openstack/nova master: Update etherpad url https://review.openstack.org/457198 | |
| 10:28:41 | bhagyashris | Hi all, I have one doubt, Actually I have tested one scenario: Steps: 1. I have created 70 instances parallely using (nova boot --flavor <id> --image <id> --max-count 70 test) all instances are created successfully with status active but only 64 instances allocations entries are created in allocations table of placement database. | |
| 10:28:54 | bhagyashris | I have tried to debug and check the code but haven't found Why only 64 instances allocation details are created even if the 70 instances are created successfully. | |
| 10:29:57 | openstackgerrit | Huan Xie proposed openstack/nova master: XenAPI: Don't use nicira-iface-id for XenServer VIF https://review.openstack.org/451641 | |
| 10:31:23 | cdent | bhagyashris: interesting. how many compute-nodes do you have? are there any messages of failed attempts to write the allocations in the logs? are you checking for the 64 allocation records instantly? is it the same a few minutes later? | |
| 10:31:29 | bhagyashris | 64 instance allocation entries means total 192 rows in allocation table (for each instance 3 records are created [vcpus, disk_gb, memory_mb]) | |
| 10:32:56 | cdent | (I was using "records" more generally than database rows) | |
| 10:32:59 | bhagyashris | cdent: I have one compute node. No there is no any failed message log | |
| 10:33:31 | cdent | can you share the logs for nova-compute and the placement-api? | |
| 10:34:12 | bhagyashris | cdent: I have checked after all the instances are created successfully | |
| 10:34:52 | openstackgerrit | Huan Xie proposed openstack/nova master: WIP: XenAPI use os-xenapi V2 in nova https://review.openstack.org/453493 | |
| 10:34:55 | bhagyashris | cdent: And one more scenario I want to discuss, | |
| 10:35:25 | cdent | the resource tracker is "self-healing" with regard to allocations: if some allocations don't get written immediately, they should be written later | |
| 10:35:54 | cdent | what you might be seeing is that the "generation" value is conflicting during some of the writes, so they fail, and then they will be correct a bit later | |
| 10:36:29 | cdent | (in fact for 70 instances all on the same node that's quite likely) | |
| 10:37:33 | bhagyashris | cdent:yeah I have checked the database entries some times later also after creating the instance but still it showing me the allocations entries for 64 instances only | |
| 10:38:34 | cdent | bhagyashris: if we can see the logs for around that time it would be great, but it definitely sounds like there's a blog | |
| 10:40:11 | cdent | bhagyashris: what's the other scenario? | |
| 10:40:15 | bhagyashris | cdent: ok. | |
| 10:42:23 | bhagyashris | cdent: scenario 2: I have created 64 instances parallely using (nova boot --flavor <id> --image <id> --max-count 64 test) then I have tried to create one more instance then in that case the instance goes into the error state by giving the Novalid host found at n-cond side and RetryFilter is returning the zero host at n-sch | |
| 10:45:01 | cdent | bhagyashris: what's your inventory for that compute-node? | |
| 10:47:06 | bhagyashris | cdent: the inventories are available. I mean the request resource are available on my compute-node | |
| 10:47:08 | cdent | my guess is that you're running out of inventory in both cases, but in the first scenario the availalbe inventory and the allocations are racing, allowing more nodes to be placed than ought to be possible. however I'm not sure how/why the compute-node would allow that | |
| 10:48:32 | cdent | bhagyashris: I think we're going to need to gather a lot more info to diagnose this so I think the right thing to do would be for you to create a bug report and include or attach: are you master or ocata? the inventory of the compute node. the logs for nova-compute and placement-api. | |
| 10:48:46 | cdent | because without more data I can't do anything more than guess | |
| 10:49:34 | bhagyashris | cdent: ok I will share all the info by reporting the bug. | |
| 10:49:47 | cdent | and at the moment I need to go to an appt, but am very curious about this so being able to track the bug would be great | |
| 10:49:54 | cdent | thanks | |
| 10:50:23 | bhagyashris | cdent: Thank you, for you time and info | |
| 10:50:42 | bhagyashris | cdent: Yeah sure I will share bug id too | |
| 10:50:42 | cdent | thank you for trying things out. we haven't had enough of that, so you doing so is great. | |
| 10:51:02 | bhagyashris | cdent: :) | |
| 11:02:37 | openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova master: Remove test_init_nonexist_schedulerdriver https://review.openstack.org/457210 | |
| 11:02:37 | openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova master: Remove usage of parameter enforce_type https://review.openstack.org/457188 | |
| 11:31:33 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/nova master: Add Neutron port capabilities to devspec in request https://review.openstack.org/451777 | |
| 11:52:09 | pradiprwt | hi Everyone, I am facing some issue with placement api in nova while installation og ocata version | |
| 11:52:31 | pradiprwt | Failed to retrieve filtered list of resource providers from placement API for filters {'resources': 'DISK_GB:1,MEMORY_MB:64,VCPU:1'}. Got 404: <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> | |
| 11:52:41 | pradiprwt | The requested URL /placement/resource_providers was not found on this server | |
| 12:01:58 | cdent | pradiprwt: how did you do your install? there are three mostly likley problems: one is that placement hasn't been registered in the service catalog, another is that the apache configuration is wrong in two different ways: a missing |
|
| 12:02:38 | pradiprwt | cdent : sure | |
| 12:04:56 | pradiprwt | cdent : which apache configuration file I should send, is that /etc/apache2/apache2.conf ..? | |
| 12:05:50 | cdent | pradiprwt: do you have an /etc/apache2/site-enabled/placement.conf or something similar to that? | |
| 12:07:51 | pradiprwt | cdent : /etc/apache2/sites-enabled/nova-placement-api.conf ::: http://paste.openstack.org/show/606890/ | |
| 12:08:04 | cdent | thank pradiprwt | |
| 12:08:16 | cdent | let me see if i can relocate a recent bug that is related to this | |
| 12:09:56 | openstack | bugzilla.redhat.com bug 1441208 in openstack-nova "placement api apache configuration does not align with documentation, ensure system does not work" [High,Closed: duplicate] - Assigned to nlevinki | |
| 12:09:56 | cdent | pradiprwt: see https://bugzilla.redhat.com/show_bug.cgi?id=1441208 and https://bugzilla.redhat.com/show_bug.cgi?id=1441200 | |
| 12:09:57 | openstack | bugzilla.redhat.com bug 1441200 in openstack-nova "openstack-placement-api.conf not configured properly by the openstack-nova-placement-api rpm" [Unspecified,Post] - Assigned to nlevinki | |
| 12:10:24 | cdent | pradiprwt: there's a section missing | |
| 12:12:35 | pradiprwt | cdent : which section is missing ? | |
| 12:12:39 | cdent | pradiprwt: see what I've added at the end here: http://paste.openstack.org/show/606892/ | |
| 12:12:53 | cdent | that may or may not be needed, depending on what's in your service catalog for placement | |
| 12:13:44 | pradiprwt | cdent : let me try this | |
| 12:20:29 | pradiprwt | cdent : same error I am getting after adding the section in /etc/apache2/sites-enabled/nova-placement-api.conf | |
| 12:20:56 | cdent | pradiprwt: what's in your service catalog as the endpoint for placement? | |
| 12:22:07 | cdent | pradiprwt: what document did you use to set things up? are you installing from packages or some other method? | |
| 12:23:18 | pradiprwt | cdent : I have used ocata official documentation using ubuntu 16.04 | |
| 12:24:24 | pradiprwt | https://docs.openstack.org/ocata/install-guide-ubuntu/ | |
| 12:25:19 | cdent | pradiprwt: okay, I think it's still the case that either your apache configuration is not complete, or your service catalog has the wrong information in it. so you could share your service catalog endpoint for placement, that would be useful | |
| 12:26:48 | pradiprwt | cdent : endpoint list cdent | |
| 12:27:12 | pradiprwt | http://paste.openstack.org/show/606894/ | |
| 12:30:57 | cdent | pradiprwt: there's a few different issues on the changes I provided in the paste i gave you, here' s a different version: http://paste.openstack.org/show/606895/ | |
| 12:41:33 | pradiprwt | cdent: thank, now it's working | |
| 12:41:43 | cdent | yay! | |
| 12:41:43 | pradiprwt | I am able to create instance | |
| 12:41:49 | pradiprwt | :) | |
| 12:42:23 | pradiprwt | cdent: Is there any problem with documentation ? | |
| 12:42:32 | cdent | pradiprwt: you may want to report a bug with ubuntu, because that configuration that's in their packages is not right | |
| 12:42:45 | cdent | no the docs are okay, I think, it seems the package is wrong | |
| 12:42:50 | pradiprwt | cdent : ok | |
| 12:43:37 | cdent | jamespage: are you aware of the above stuff with packaging of placement? | |
| 12:54:53 | openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova master: Remove test_init_nonexist_schedulerdriver https://review.openstack.org/457210 | |
| 12:54:53 | openstackgerrit | ChangBo Guo(gcb) proposed openstack/nova master: Remove usage of parameter enforce_type https://review.openstack.org/457188 | |
| 13:07:43 | openstack | Launchpad bug 1672625 in OpenStack Compute (nova) "Instance stuck in schedule state in Ocata release" [Undecided,New] | |
| 13:07:43 | sdague | dansmith / melwitt - when you get up and about - https://bugs.launchpad.net/nova/+bug/1672625 | |
| 13:07:58 | sdague | it looks like something missing on the cells enablement in the docs maybe? | |
| 13:17:05 | dane-fichter | anyone want to review some security code? there's a bug fix for ephemeral storage encryption that hasn't gotten a review in almost a month | |
| 13:17:10 | dane-fichter | https://review.openstack.org/#/c/446072/ | |
| 13:17:30 | dane-fichter | ^ ephemeral storage encryption is broken right now, this fixes it | |
| 13:22:43 | efried | edmondsw ^^ ? | |
| 13:24:13 | edmondsw | efried, already has a bunch of +1s... I think they're looking for some +2s | |
| 13:24:37 | efried | edmondsw Yuh, hadn't looked at it yet when I tagged ya. Just saw "security" and... | |
| 13:24:45 | edmondsw | :) | |
| 13:30:25 | ericyoung | if core reviewer wants to look over some code, I could use a second + on https://review.openstack.org/#/c/452538/ | |
| 13:30:40 | dansmith | sdague: nope, i think it's all on that doc, including the step they missed | |
| 13:30:57 | dansmith | sdague: dunno about the ubuntu docs, but the first one linked which we own has got the stuff | |