16:00:24 #startmeeting Octavia 16:00:24 Meeting started Wed Aug 24 16:00:24 2022 UTC and is due to finish in 60 minutes. The chair is gthiemonge. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:25 The meeting name has been set to 'octavia' 16:00:30 hi folks! 16:00:31 o/ 16:00:50 o/ 16:02:02 hi, how do i change the load balancer status from ERROR to ACTIVE as the loadbalancer is working fine.. the loadbalancer went in to error state because of compute node rebooting.. 16:02:44 viks__ Use the load balancer failover function. 16:02:58 +1 16:03:31 #link https://docs.openstack.org/python-octaviaclient/latest/cli/index.html#loadbalancer-failover 16:04:51 Compute node rebooting (or the VM for that matter) can mean some of the cryptographic information was lost as it's in an encrypted RAM drive inside the amphora. 16:06:52 #topic Announcements 16:07:05 ** OpenStack Zed schedule 16:07:16 (same as last week ;)) 16:07:46 next week is the Feature Freeze/Final release for client libraries! 16:07:48 Next week is the deadline for octavia-lib 16:07:57 I see good progress on the reviews! 16:08:05 at least: activity ;-) 16:08:12 #link https://releases.openstack.org/zed/schedule.html 16:08:24 I don't think we have anything expected for zed left in octavia-lib 16:08:56 ok, and we released octavia-lib 3.1.0 last week 16:09:40 johnsom: thanks.. for failover we need active-standby setup with anti-affinity right? with single node, it goes in to ` Pending Update ` and then back to `error` 16:10:10 ... which means that I was able to finish the failover circuit breaker change 16:10:32 viks__ No, it works with standalone as well. It should not go back to error. What do the logs say for why it is going back to error? 16:11:14 hey you might have seen that I have created a lot of backports 16:11:38 as the fixes for Zed are getting merged, I'm backporting them (I don't want to miss a fix) 16:11:44 Standalone failover just has a short amount of downtime during a failover, were active/standby does not. 16:11:51 reviewing these backports is not really urgent 16:12:09 +1 to doing backports as things merge. 16:12:14 I think when Zed is released and those backports are merged, we will deliver a new stable releases 16:12:36 also worth mentioning: we had no bare rechecks recently. Our statistics are clean ATM 16:12:42 https://etherpad.opendev.org/p/recheck-weekly-summary 16:12:46 lol 16:13:19 the script is broken? 16:13:27 probably 16:13:50 well, this is great ;-) 16:14:00 I mean 0 bare recheck 16:14:24 Can we have a badge? 16:14:41 "7 days since the last bare recheck" 16:15:34 johnsom: ` ComputeBuildException: Failed to build compute instance due to: {u'message': u'No valid host was found. There are not enough hosts available` 16:16:13 viks__ Ok, so your nova is broken and we can't build a new amphora. You need to fix nova, then try again. 16:17:44 but the amphora instances are running fine, and the requests sent to load balancer being sent to the instances behind it.. i get appropriate response 16:18:43 only `provision state` is ERROR.. i.e. as mentioned in https://docs.openstack.org/octavia/latest/admin/guides/operator-maintenance.html#monitoring-load-balancers 16:18:43 #topic CI Status 16:18:48 nothing to report here 16:18:54 viks__ Yep, that can be the case. Operating status will reflect that. But it may be out of sync with the control plane. 16:19:07 `provisioning status of ERROR, but an operating status of ONLINE` 16:19:50 is there any way to bring back the operating status in sysnc with control plane in such scenarios? 16:20:00 viks__ Yep. Given the error from nova "No valid host was found.", you probably can't create new load balancers or VMs either. 16:20:21 viks__ Yes, the failover mechanism is for correcting this issue. 16:22:13 johnsom: ok... i'll will further investigate whats happening.. thanks for the quick response 16:23:00 #topic Brief progress reports / bugs needing review 16:23:23 I have been focused on reviews. Making my way through the subnet/multi-vip patch chain 16:23:32 I did many reviews... that's all :D 16:23:41 Big and complex patches.... 16:24:49 the octavia priority review list lists two notifications implementations in addition to the one that was merged recently. shouldn't we remove them from the list? 16:25:07 Yeah, we should probably abandon those patches too 16:25:28 yeah they are outdated and unmaintained 16:25:48 Experimental multi-az support is also on the list, I know that it is not ready too 16:27:07 "Use vip_subnet_selection_tag to select subnets" will conflict with the multi-subnets on members and multi-vip patches :/ 16:30:26 #topic Open Discussion 16:30:52 FYI, from the discuss mailing list, there is more info on this story: https://storyboard.openstack.org/#!/story/2007619 16:31:13 There appears to be an issue with one of the "special" token types and accessing barbican. 16:31:55 * johnsom notes this is not related to the scoped tokens RBAC stuff 16:32:11 johnsom: is it related to application credentials? 16:32:39 "trust-scoped" tokens. I'm not 100% sure I even know what that is 16:33:08 Kong fixed it for neutron in the past. 16:34:52 yeah we should ask for a simple reproducer (or advices for the conf) 16:36:02 I just wanted to bring it up at the meeting so someone might be able to pick this up. 16:36:51 johnsom: I will follow up in the story 16:37:01 Awesome, thanks 16:39:52 anything else? 16:40:07 not from me 16:40:08 I don't think I have anything else 16:42:58 ok thank you guys 16:43:04 #endmeeting