| tychicus | did you initialize vault? | 00:00 |
|---|---|---|
| tychicus | I ran into this when testing the new 20.05 stable bundle | 00:01 |
| tychicus | there are 2 sets of instructions that I needed to get things moving forward | 00:01 |
| tychicus | https://docs.openstack.org/project-deploy-guide/charm-deployment-guide/latest/app-vault.html | 00:02 |
| tychicus | first initialize and install vault | 00:02 |
| tychicus | https://docs.openstack.org/project-deploy-guide/charm-deployment-guide/latest/app-certificate-management.html | 00:02 |
| tychicus | second retrieve and sign the csr | 00:02 |
| wallyworld | thumper: i give up, 2.7 landing is broken, here's a forward port of the 2.7 PR directly to 2.8-rc3 (with legacy lease stuff removed) https://github.com/juju/juju/pull/11641 | 01:00 |
| wallyworld | or hpidcock or kelvinliu ^^^^^ | 01:06 |
| wallyworld | kust got to get 2.8-rc unblocked | 01:06 |
| kelvinliu | looking | 01:07 |
| kelvinliu | approved since the original one has been approved already wallyworld | 01:10 |
| wallyworld | ty | 01:10 |
| kelvinliu | nws | 01:11 |
| timClicks | I'm receiving a "no matching agent binaries available" message when deploying a focal workload to AWS | 01:58 |
| wallyworld | juju version? | 02:04 |
| timClicks | this is JAAS (2.6.10), client is Juju 2.8-rc2 | 02:05 |
| thumper | timClicks: jaas doesn't know about focal | 02:05 |
| thumper | that is 2.7 | 02:05 |
| thumper | wallyworld: where are we at with the landings? | 02:05 |
| wallyworld | gave up on 2.7 and forward ported the last PR directly to 2.8-rc, waiting for CI now | 02:06 |
| thumper | wallyworld: can I try to merge my 2.8 branch? | 02:06 |
| wallyworld | sure, i'm just landing a 2.8rc->2.8 port and then i was going to merge yours | 02:07 |
| thumper | all good, if you want to do the mergy thing then that's all good | 02:07 |
| wallyworld | i can | 02:08 |
| thumper | looks like I'll be helping with a customer issue | 02:08 |
| wallyworld | what juju version? | 02:09 |
| thumper | don't know yet | 02:09 |
| hpidcock | forward port to develop https://github.com/juju/juju/pull/11642 | 03:02 |
| suke | hi Guys, | 03:03 |
| wallyworld | hpidcock: did you pick up the just landed pr from tim? | 03:05 |
| hpidcock | wallyworld: let me check | 03:06 |
| wallyworld | hpidcock: also, GetByHardwareAddress isn't used | 03:07 |
| wallyworld | i suspect it's been replaced by Filter whichis whwere the conflict was | 03:07 |
| hpidcock | GetByHardwareAddress was only added 10 days ago | 03:10 |
| wallyworld | hpidcock: ah, i suspect it's for the work to filter out machine local addresses, and there's a PR which will use it | 03:14 |
| hpidcock | wallyworld: yep, can always delete it later. I'd rather not break manadart's WIP if he has any using it | 03:15 |
| wallyworld | yup | 03:15 |
| wallyworld | hpidcock: the changes look as expected to me based on working with the other branches of late | 03:18 |
| hpidcock | wallyworld: awesome thanks, just waiting for a green run then I'll merge | 03:19 |
| wallyworld | i was 1/2 way thriugh doing it myself then i saw your pr :-) | 03:19 |
| hpidcock | wallyworld: ahah | 03:19 |
| wallyworld | was looking at the nic conflict | 03:20 |
| hpidcock | sorry | 03:20 |
| wallyworld | all good, you saved me a lot6 of typing | 03:20 |
| thumper | https://github.com/juju/juju/pull/11644 | 04:17 |
| thumper | the last of the uniter package level loggers | 04:18 |
| thumper | phew | 04:18 |
| manadart | wallyworld, hpidcock: GetByHardwareAddress was indeed a utility added for a patch in progress. | 04:25 |
| manadart | wallyworld: This is the fix for network_get and local-machine addresses: https://github.com/juju/juju/pull/11638 | 04:26 |
| manadart | Just have to put together some QA. | 04:26 |
| wallyworld | gr8 ty | 04:27 |
| Eryn_1983_FL | tychicus: he got some handful of keys | 04:29 |
| Eryn_1983_FL | i think he did | 04:29 |
| * thumper sighs | 05:30 | |
| thumper | more problems | 05:30 |
| thumper | but yay, users?? | 05:30 |
| wallyworld | thumper: anything serious? | 05:35 |
| thumper | wallyworld: nothing blocking | 05:35 |
| thumper | but enough that we'll want to include a patch in 2.7.7 | 05:35 |
| wallyworld | got a bug? | 05:35 |
| thumper | https://bugs.launchpad.net/juju/+bug/1881242 | 05:36 |
| mup | Bug #1881242: Missing error check results in panic - apiserver <apiserver> <panic> <juju:Triaged> <https://launchpad.net/bugs/1881242> | 05:36 |
| thumper | the fix for this bug is trivial... add an error handling | 05:36 |
| thumper | the other problem, for which we are also getting a bug, is how did it get into this status | 05:36 |
| thumper | state | 05:36 |
| wallyworld | sigh | 05:37 |
| thumper | wallyworld: did you see my PR above? | 05:39 |
| thumper | wallyworld: nm, hpidcock reviewed it for me | 05:40 |
| thumper | thanks hpidcock | 05:40 |
| wallyworld | that's why i didn't look :-) | 05:40 |
| wallyworld | already +1 | 05:40 |
| thumper | hey... https://jenkins.juju.canonical.com/job/github-make-check-juju/5933/consoleText | 05:41 |
| thumper | look for "oops" | 05:42 |
| thumper | api/uniter fails | 05:42 |
| thumper | but I don't see a filure in the output? | 05:42 |
| hpidcock | thumper: might be teardown failure? | 05:45 |
| thumper | succeeds locally | 05:46 |
| thumper | and I only see 118 passed and 1 skipped | 05:46 |
| thumper | hpidcock: could be | 05:46 |
| thumper | wallyworld: I see the email you just forwarded to me | 05:50 |
| thumper | but I don't see the original | 05:50 |
| wallyworld | thumper: yeah, go figure, maybe the email lists are stuck | 05:50 |
| thumper | could be | 05:50 |
| thumper | wallyworld: you message to the list has come through now | 06:34 |
| wallyworld | so it has | 06:35 |
| Laney | hey, quick bit of help needed hopefully. i've got a controller running on lxd locally, but juju has somehow got the wrong idea about what its ip addresses are (maybe the container got restarted?). how can I get it to learn the right ones? | 09:00 |
| Laney | (made a topic on discourse) | 09:15 |
| achilleasa | Laney: you can point the cli to the right IP by editing the endpoint values in ~/.local/share/juju/controllers.yaml for your controller (as always backup first before editing ;-) | 09:17 |
| Laney | achilleasa: wtf, that has the right addresses in it | 09:18 |
| Laney | uh | 09:18 |
| Laney | there's some duplication here, how did that happen | 09:18 |
| achilleasa | Laney: you mean you have the same IP twice in the endpoint list? Not sure how that could happen but it should not have any impact in connection attempts | 09:20 |
| Laney | no I had the same controller in there twice, once right and once wrong | 09:20 |
| Laney | but deleted the wrong one now and 'juju status' works again | 09:20 |
| Laney | 'juju ssh' doesn't though, it seems to be trying to connect to my public IP instead of the container's IP ... | 09:21 |
| Laney | works with --proxy but not without | 09:24 |
| achilleasa | manadart: can you take a look at https://github.com/juju/juju/pull/11645? | 11:39 |
| manadart | achilleasa: Approved it. | 11:45 |
| achilleasa | manadart: do we rewrite the agent config files when upgrading? | 12:04 |
| manadart | achilleasa: Yes, we have to in order to set the version. | 12:05 |
| Eryn_1983_FL | hi | 12:14 |
| tychicus | with the nova-cloud-controller charm, is there bit of additional config that needs to be performed to get console-access-protocol to use TLS? | 19:14 |
| tychicus | I ensured that the vault relation is there juju add-relation nova-cloud-controller:certificates vault:certificates | 19:15 |
| tychicus | but openssl s_client -connect shows that TLS is not enabled | 19:17 |
| tychicus | If I switch from https to http the page renders with an error "Something went wrong, connection is closed " | 19:18 |
| pmatulis | tychicus, are you talking about the dashboard? | 19:19 |
| tychicus | yes, the console in the dashboard | 19:20 |
| tychicus | dashboard loads with https no problem | 19:20 |
| tychicus | in previous releases to enable ssl you had to supply a console-ssl-cert to nova-cloud-controller | 19:22 |
| tychicus | but console-ssl-cert was removed in the 19.07 charm release | 19:23 |
| tychicus | Please use ssl_cert configuration option or the vault certificates relation | 19:24 |
| pmatulis | tychicus, where do you see that? | 19:25 |
| tychicus | https://jaas.ai/nova-cloud-controller/345#charm-config-console-ssl-cert | 19:26 |
| pmatulis | tychicus, thanks | 19:33 |
| tychicus | pmatulis: np, just trying to figure out if I missed something trivial in my config | 19:34 |
| pmatulis | tychicus, what about the value of 'console-access-protocol'? | 19:35 |
| tychicus | pmatulis: uju config nova-cloud-controller console-access-protocol=novnc | 19:35 |
| tychicus | s /uju/juju | 19:36 |
| pmatulis | tychicus, k | 19:36 |
| pmatulis | tychicus, well, based on what you've said you've done and based on the documentation it should work. can you open a bug on nova-cloud-controller? | 19:37 |
| tychicus | sure | 19:37 |
| pmatulis | tychicus, did you try logging out and back in to the dashboard after having made that config change (to 'novnc')? | 20:31 |
| tychicus | I did | 20:32 |
| pmatulis | k | 20:33 |
| pmatulis | well, like i recommend: https://bugs.launchpad.net/charm-nova-cloud-controller/+filebug | 20:34 |
| tychicus | I'm working on the bug report now, but I need to do a little more digging | 20:34 |
| tychicus | it looks like novnc is not proxied by apache, all of the stuff proxied by apache is fine | 20:35 |
| tychicus | looks like the issue is that there is an issue with the nova.console.websocketproxy | 21:13 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!