| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-06 | |||
| 07:04:44 | litao | ' | |
| 07:04:48 | alex_xu | litao: no, I mean the meanning of 'updated_at' is for that | |
| 07:05:23 | litao | alex_xu: but the output of 'nova service-list' will display the column | |
| 07:05:35 | litao | alex_xu: updated_at | |
| 07:06:11 | alex_xu | litao: yes, 'updated_at' means 'the timestamp when db record of service updated' | |
| 07:07:18 | litao | so i think the value should be updated regardless of db or mc driver | |
| 07:08:46 | litao | alex_xu: If it retains for a long time not to update, it will give us a doubt that is the service really up? | |
| 07:09:14 | alex_xu | litao: the patch https://review.openstack.org/#/c/163060/ separate the db record update time and the heartbeat update time. After that patch 'updated_at' means db record time. Maybe we can have another field for heartbeat timestamp, but i'm not sure | |
| 07:09:39 | alex_xu | litao: yes, but | |
| 07:09:51 | alex_xu | 'updated_at' isn't for determine whether the service is up | |
| 07:10:54 | alex_xu | 'last_seen_at' and your change for mc is for determine that, but that isn't expose in the API | |
| 07:11:27 | litao | alex_xu: but i think it should be same as db drvier to be updated since it presents to us | |
| 07:11:59 | litao | alex_xu: or remove it when using mc driver | |
| 07:12:01 | alex_xu | litao: db driver didn't update 'updated_at' also | |
| 07:13:03 | litao | alex_xu: it does | |
| 07:13:12 | alex_xu | litao: there is field 'State' for determine whether service is up, why you need to lookup another timestamp? | |
| 07:14:05 | litao | alex_xu: I mean that it just a display, not to determine the state of a service | |
| 07:15:12 | litao | alex_xu: I just want to display the same output of 'nova service-list' when using mc driver | |
| 07:20:10 | alex_xu | litao: emm...I see now, it will update. sorry, I missed that. But it still didn't equal to the your change for the mc driver I think | |
| 07:21:10 | alex_xu | litao: for db driver, 'updated_at' mixes db record update and heartbeat update. | |
| 07:24:36 | litao | alex_xu: yes, so what do you think to deal it? | |
| 07:24:46 | alex_xu | litao: I just comment on the patch | |
| 07:58:35 | openstackgerrit | Sergey Nikitin proposed openstack/nova-specs master: PCI NUMA Policies https://review.openstack.org/361140 | |
| 08:23:52 | openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs master: Repropose tag-instance-when-boot https://review.openstack.org/415315 | |
| 08:44:32 | openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs master: Send out notifications when instance tags chaned https://review.openstack.org/429569 | |
| 08:53:03 | openstackgerrit | Zhenyu Zheng proposed openstack/nova-specs master: Extend tag length limit to 80 https://review.openstack.org/423112 | |
| 09:01:34 | litao | alex_xu: thanks | |
| 09:12:54 | alex_xu | litao: np | |
| 10:24:21 | FransUrbo | I'm trying to figure out why my second primary servers services won't come up. They're listed as 'down' with OSC, but the 'service:updated_at' column have been updated very recently - https://gist.github.com/FransUrbo/a24e99a331c2a252ecdc76a5c6f12b20 (that's for Cinder, but Nova shows the same problem). | |
| 10:26:19 | FransUrbo | https://gist.github.com/FransUrbo/e11eb9d9311790c63729f16e108dbb1f (for nova) | |
| 10:29:38 | BobBall | johnthetubaguy: Are you happy with https://review.openstack.org/#/c/424428/ ? Clearly the change missed RC1 but I'd love to get it in RC2 if we can get reviews soon | |
| 10:30:10 | johnthetubaguy | BobBall: we spoke about it being eligible for RC2 on friday evening | |
| 10:30:35 | johnthetubaguy | BobBall: got my head in a spec right now, once my head is out of that, I will take a peak at it | |
| 10:30:40 | BobBall | Yeah - I spotted the change in the etherpad too. Awesome, thanks! | |
| 10:31:09 | johnthetubaguy | no worries, it would be a shame for that bug to stick around another release | |
| 10:32:18 | BobBall | Indeed. Getting the fix in is my priority ATM so let me know what I need to do. | |
| 10:50:39 | openstackgerrit | Bob Ball proposed openstack/nova master: XenAPI: fix memory usage miscalculation https://review.openstack.org/394857 | |
| 11:05:24 | openstackgerrit | sahid proposed openstack/nova master: libvirt: update the active migrations DS to support migration result https://review.openstack.org/409260 | |
| 11:05:25 | openstackgerrit | sahid proposed openstack/nova master: libvirt: ensure live-migration success https://review.openstack.org/409261 | |
| 11:39:32 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: api-ref: Fix network_label parameter type https://review.openstack.org/429659 | |
| 11:53:32 | artom_ | In the api samples tests, if the response template doesn't exit, it's created with the response data | |
| 12:45:16 | openstackgerrit | John Garbutt proposed openstack/nova-specs master: WIP: Spec idea on the future of quotas for Nova https://review.openstack.org/429678 | |
| 12:58:32 | johnthetubaguy | BobBall: looks good to me | |
| 13:04:58 | BobBall | Thanks johnthetubaguy | |
| 13:45:18 | openstackgerrit | Timofey Durakov proposed openstack/nova master: abort_live_migration remove redundant check https://review.openstack.org/304621 | |
| 13:56:08 | mriedem | BobBall: i'm in a timewarp http://dd6b71949550285df7dc-dda4e480e005aaa13ec303551d2d8155.r49.cf1.rackcdn.com/76/429476/1/check/dsvm-tempest-neutron-network/00891c1//index.html | |
| 13:56:15 | mriedem | clicking logs takes me to a dir that just takes me back | |
| 13:56:46 | BobBall | mriedem: Did you take a jump to the left? And then a step to the right? | |
| 13:57:15 | mriedem | i forgot those steps | |
| 13:57:55 | bauzas | folks, FYI, will mostly be afk this week because I'm organizing a local conf named http://snowcamp.io/ | |
| 13:58:04 | bauzas | with ttx keynoting :p | |
| 13:58:07 | BobBall | mriedem: It looks like the rest of the logs didn't upload curiously | |
| 13:58:32 | BobBall | mriedem: Received disconnect from 192.168.33.2: 2: Packet corrupt <-- or there were no other logs. | |
| 13:58:52 | BobBall | mriedem: that test run didn't even get to the point of running devstack | |
| 13:59:04 | edleafe- | Scheduler subteam meeting in 1 minute over in #openstack-meeting-alt | |
| 13:59:27 | mriedem | BobBall: ok i just rechecked it | |
| 13:59:53 | BobBall | mriedem: We do put "run_tests.log" in both directories... which is perhaps unhelpful. | |
| 14:00:48 | ttx | \o/ and fresh snow | |
| 14:02:58 | bauzas | ttx: zactly, 30cm waiting for you ;) | |
| 14:05:30 | openstackgerrit | Timofey Durakov proposed openstack/nova master: nfs fix for xenial images https://review.openstack.org/329466 | |
| 14:06:09 | kashyap | sdague: Heya, random ping -- someone (a Kernel developer, actually) at FOSDEM asked me yesterday: Why DevStack, why not...just use 'Gentoo'. They were serious :-) | |
| 14:06:33 | sdague | kashyap: :) | |
| 14:06:45 | sdague | sounds like they never actually looked into the issue | |
| 14:07:03 | openstackgerrit | Stephen Finucane proposed openstack/nova master: console: Move proxies to 'console/proxy' https://review.openstack.org/408192 | |
| 14:07:12 | kashyap | sdague: Yeah, right. But they were not really mean or anything, just thinking out loud | |
| 14:07:36 | kashyap | sdague: I explained them, also the reason why it's shell script (that a human / admin / operator should be able to read it, etc). | |
| 14:10:02 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Deprecate two more nova-net options https://review.openstack.org/412390 | |
| 14:23:34 | mriedem | superdan: first, where is dansmith | |
| 14:23:43 | mriedem | superdan: second, alex_xu and i have comments in https://review.openstack.org/#/c/428824/ | |
| 14:32:02 | dansmith | mriedem: okay coffee then that | |
| 14:34:22 | openstack | Launchpad bug 1659391 in OpenStack Compute (nova) "Server list API does not show scheduled servers that are not assigned to any cell" [Undecided,New] | |
| 14:34:22 | mriedem | manila also reported a bug that the change for multi-cell instance listing broke them https://bugs.launchpad.net/nova/+bug/1659391 but so far i'm not able to find a job that failed due to that to debug it, | |
| 14:34:59 | mriedem | plus the description is overly confusing with references to cells being disabled and default cells, which really throws me off | |
| 14:38:17 | dansmith | hmm, okay | |
| 14:39:01 | dansmith | mriedem: well the description says that after that change, nova only returns instances assigned to cells | |
| 14:39:03 | dansmith | which is 100% true :) | |
| 14:40:02 | mriedem | yeah i left some more questions, it's not clear to me if the server they are creating is done before simple_cell_setup or not | |
| 14:40:39 | dansmith | like before an upgrade you mean? | |
| 14:41:41 | mriedem | "It can be seen when cell feature and services are disabled and default cell hasn't been created yet. In this case all Nova VMs hang in scheduled state and cannot be viewed." | |
| 14:41:45 | mriedem | i was just asking wtf that meant | |
| 14:41:57 | mriedem | but it looks like simple_cell_setup is run first http://logs.openstack.org/37/426737/12/check/gate-manila-tempest-dsvm-generic-no-share-servers-ubuntu-xenial-nv/064f0da/logs/devstacklog.txt.gz#_2017-02-02_13_41_26_166 | |
| 14:42:07 | mriedem | and then they create a server later http://logs.openstack.org/37/426737/12/check/gate-manila-tempest-dsvm-generic-no-share-servers-ubuntu-xenial-nv/064f0da/logs/devstacklog.txt.gz#_2017-02-02_13_45_01_035 | |
| 14:42:30 | dansmith | well, the first part of that sentence isn't valid either | |
| 14:42:56 | mriedem | i think they mean cells v1 | |
| 14:43:00 | dansmith | this is one of the reasons I wish we had used a different name than cells | |
| 14:43:03 | mriedem | but yeah that totally threw me off when first reading this | |
| 14:43:08 | dansmith | there's a lot of confusion that has stemmed from that decision | |
| 14:43:50 | mriedem | having said all that, the whole 'what don't you get about this bug' attitude in the report makes me kind of not care either | |
| 14:44:12 | mriedem | unless it's not unique to them | |
| 14:44:22 | dansmith | yeah | |
| 14:44:52 | dansmith | the change they're referencing has been in for like three weeks now right? | |
| 14:44:54 | mriedem | maybe i'll just revert their workaround and see if it fails on the same thing again and then debug from there | |
| 14:45:08 | mriedem | yeah, landed on 1/24 and they worked around it on 1/28 it looks like | |
| 14:45:17 | dansmith | doubt anyone else has had their gate completely shutdown for three weeks | |
| 14:45:27 | sfinucan | There's probably a much better solution, but hacky hack for opening a Python module path in vim https://that.guru/blog/open-python-paths-with-vim/ | |
| 14:45:30 | dansmith | and just before rc2 they bring it up.. nice | |
| 14:47:16 | mriedem | the bug was actually reported awhile ago | |
| 14:47:27 | mriedem | 1/25 | |
| 14:47:32 | mriedem | the day after your change landed https://bugs.launchpad.net/nova/+bug/1659391 | |
| 14:47:33 | openstack | Launchpad bug 1659391 in OpenStack Compute (nova) "Server list API does not show scheduled servers that are not assigned to any cell" [Undecided,New] | |