| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-07-19 | |||
| 17:14:08 | bauzas | mriedem: since we didn't added a relnote for https://review.openstack.org/#/c/416521/63 do you think we should at least add a note saying that the 'reserved' QS is no longer accepted in Pike as per https://review.openstack.org/#/c/446242/41 ? | |
| 17:14:16 | bauzas | dansmith: melwitt: ^ | |
| 17:14:37 | mriedem | bauzas: it is accepted | |
| 17:14:39 | mriedem | it's just ignored | |
| 17:15:00 | mriedem | the end cloud user isn't going to read the release notes, they are going to read the api reference | |
| 17:15:06 | mriedem | and the api doesn't change | |
| 17:15:12 | bauzas | mriedem: sure, I meant telling people that if they use it, it won't no longer be something we check | |
| 17:15:22 | bauzas | meh, good point | |
| 17:15:23 | mriedem | no, that's just noise imo | |
| 17:15:31 | bauzas | it's more an API doc | |
| 17:15:32 | dansmith | and we're not reserving so ignoring it is reasonable right? | |
| 17:15:45 | mriedem | yeah this is all internal tomfoolery | |
| 17:16:02 | mriedem | as i noted in there, i'll propose removing that query parameter with a microversion in queens | |
| 17:16:03 | dansmith | it's noise, IMHO | |
| 17:16:03 | dansmith | reserved amounts are transient | |
| 17:16:03 | dansmith | if we improved boot performance, you'd see the same behavior | |
| 17:19:33 | bauzas | mriedem: dansmith: can't find the API reference for that API endpoint FWIW | |
| 17:19:39 | bauzas | am I blind ? | |
| 17:19:48 | mriedem | limits | |
| 17:19:56 | mriedem | another way to think of that is if the cloud used the noop quota driver, | |
| 17:19:57 | melwitt | the used limits API is read only, it's just showing "reserved" usage | |
| 17:20:02 | mriedem | there wouldn't be reservations | |
| 17:20:17 | mriedem | https://developer.openstack.org/api-ref/compute/?expanded=show-rate-and-absolute-limits-detail | |
| 17:20:33 | bauzas | https://developer.openstack.org/api-ref/compute/#show-rate-and-absolute-limits | |
| 17:20:34 | mriedem | https://developer.openstack.org/api-ref/compute/?expanded=show-rate-and-absolute-limits-detail#limits-limits | |
| 17:20:37 | bauzas | yeah | |
| 17:20:57 | bauzas | so my question is, should we take the opportunity to remove 'reserved' from the API doc $ | |
| 17:20:59 | bauzas | ^ | |
| 17:21:19 | mriedem | no | |
| 17:21:30 | mriedem | because the api doc is true for liberty deployments | |
| 17:21:37 | mriedem | which aren't using counting quotas | |
| 17:21:41 | melwitt | yeah, like I thought that would require a microversion to just remove it from the API | |
| 17:21:46 | bauzas | oh right | |
| 17:21:52 | bauzas | we would need to microversion | |
| 17:21:59 | melwitt | since we're not reserving, "reserved" is 0 technically | |
| 17:22:01 | mriedem | jesus yes that's what i said in the patch itself | |
| 17:22:05 | melwitt | lol | |
| 17:22:29 | jaypipes | mriedem: done. | |
| 17:22:40 | mriedem | https://review.openstack.org/#/c/446242/41/nova/api/openstack/compute/used_limits.py@a65 | |
| 17:23:07 | bauzas | argh, jaypipes beated me up | |
| 17:23:40 | bauzas | mriedem: melwitt: for my knowledge, when did we stopped using that flag? | |
| 17:23:49 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/nova master: Read Neutron port 'binding_profile' during boot https://review.openstack.org/449257 | |
| 17:23:50 | mriedem | https://review.openstack.org/#/c/416521/ | |
| 17:23:53 | mriedem | in there ^ | |
| 17:24:02 | mriedem | the last reservable resource was instances, | |
| 17:24:13 | mriedem | and that change changes instnaces from a reservable resource to a countable resource | |
| 17:24:15 | melwitt | I don't understand how we're not using it, we're just correctly saying it's zero | |
| 17:24:17 | mriedem | hence reserved is no longer used | |
| 17:24:33 | mriedem | melwitt: i think he's asking which change made reserved vestigial | |
| 17:24:37 | melwitt | instead of letting it potentially show old unused reservations | |
| 17:24:41 | mriedem | it's this gd beast https://review.openstack.org/#/c/416521/63/nova/quota.py | |
| 17:24:41 | openstackgerrit | Merged openstack/nova master: Correct the description of 'disable-log-reason' api-ref https://review.openstack.org/480209 | |
| 17:24:43 | melwitt | oh | |
| 17:25:17 | bauzas | yeah just saw https://review.openstack.org/#/c/416521/63/nova/quota.py@139 | |
| 17:25:43 | bauzas | melwitt: yeah, what mriedem said | |
| 17:26:02 | melwitt | I guess we could leave it there and it would still be 0 because in nova/quota.py reserved is hardcoded to 0 | |
| 17:26:03 | bauzas | melwitt: I just wondered when we began ignoring that API flag | |
| 17:26:25 | bauzas | I'm definitely not a quota expert :( | |
| 17:27:10 | melwitt | with each resource change from ReservableResource -> CountableResource in nova/quota.py we return reserved=0 for that resource since reserving is no longer possible | |
| 17:28:54 | bauzas | anyway, time to bail out \o | |
| 17:31:09 | melwitt | bah, counting instances failed in the gate on yet another unrelated func test "Authentication error" | |
| 17:31:28 | dansmith | well if someting in front of it fails that test it'll get another go | |
| 17:31:51 | openstackgerrit | Merged openstack/nova master: Update URL home-page in documents according to document migration https://review.openstack.org/483757 | |
| 17:32:02 | melwitt | oooh | |
| 17:32:58 | openstackgerrit | Merged openstack/nova master: Do not mention that tags are case sensitive in docs https://review.openstack.org/484957 | |
| 17:33:42 | openstackgerrit | Merged openstack/nova master: XenAPI: Fix ValueError in test_slave_asks_master_to_add_slave_to_pool https://review.openstack.org/484201 | |
| 17:34:28 | openstackgerrit | Merged openstack/nova master: Convert HostState.limits['numa_topology'] to primitive https://review.openstack.org/484439 | |
| 17:35:13 | openstackgerrit | Merged openstack/nova master: api-ref: fix max_version for deprecated os-quota-class-sets parameters https://review.openstack.org/484952 | |
| 17:35:42 | dansmith | okay not likely now :) | |
| 17:39:38 | mriedem | at least your patch got into the gate, | |
| 17:39:45 | mriedem | i can't get https://review.openstack.org/#/c/464280/ out of check queue purgatory | |
| 17:40:15 | melwitt | true. | |
| 17:45:41 | bbansal | Hello All, anyone here can help me : how can I disable genration of n-dhcp logs ? | |
| 17:49:03 | efried | bbansal symlink it to /dev/null? | |
| 17:49:17 | efried | bbansal Set the log level to CRITICAL? | |
| 17:50:54 | bbansal | I have devstack running as a part of CI. It generates n-dhcp log and its a symlink. At the end of the process, when we are trying to copy it to our ftp server, it fails because there is no such file there | |
| 17:51:33 | mriedem | you mean q-dhcp? | |
| 17:51:53 | bbansal | i am looking for n-dhcp | |
| 17:51:58 | bbansal | q-dhcp logs are fine | |
| 17:52:04 | mriedem | there is no n-dhcp | |
| 17:52:15 | bbansal | n-dhcp.log -> n-dhcp.log.2017-07-18-233804 | |
| 17:52:21 | bbansal | it generates like this | |
| 17:53:01 | bbansal | i can see these logs in my logs directory | |
| 17:53:53 | openstackgerrit | Ed Leafe proposed openstack/nova master: WIP - Migrate Ironic Flavors https://review.openstack.org/484949 | |
| 17:54:21 | bbansal | mriedem: any insights> | |
| 17:54:22 | bbansal | ? | |
| 17:55:14 | mriedem | there is no n-dhcp in devstack | |
| 17:55:15 | mriedem | on master | |
| 17:55:34 | mriedem | oh hold on | |
| 17:56:01 | bbansal | sure | |
| 17:56:12 | mriedem | i see it | |
| 17:56:36 | mriedem | https://github.com/openstack-dev/devstack/blob/master/lib/nova#L544-L554 | |
| 17:57:04 | mriedem | that should only be for running nova-network i think | |
| 17:57:06 | mriedem | sdague: ^? | |
| 17:57:27 | bbansal | yeah and we dont need these and currently they are failing our builds somehow even if all tempest cases are passing | |
| 17:58:19 | mriedem | yeah it's a bug in devstack, it shouldn't be running that if you're using neutron | |
| 17:59:20 | mriedem | i'll push a change | |
| 18:00:23 | sdague | mriedem: looking | |
| 18:01:27 | bbansal | ok thanks | |
| 18:02:14 | sdague | mriedem: yeh, probably. It doesn't hurt anything because it's a zero length file | |