* thumper is off to physio in 1.5 hours to see if he has broken his hand | 00:25 | |
babbageclunk | :( | 00:26 |
---|---|---|
anastasiamac | babbageclunk: my thoughts so far... https://github.com/juju/juju-restore/pull/3 | 00:53 |
babbageclunk | ok, will look at that after kelvinliu's one | 00:53 |
babbageclunk | (after your other one) | 00:54 |
anastasiamac | babbageclunk: no worries \o/ m not in a rush... | 00:58 |
anastasiamac | babbageclunk: also cannot seem to link juju-reestore repo in trello so will just attach the link instead of using github button | 00:59 |
wallyworld | kelvinliu: can we have a unit test for the PR? i guess you're waiting for a +1 from field? | 01:14 |
kelvinliu | wallyworld: yeah. and babbageclunk is reviewing | 01:16 |
babbageclunk | anastasiamac: maybe there's a trello user that needs access to the repo? | 01:17 |
anastasiamac | babbageclunk: dunno but attaching suited me for now :) | 01:30 |
babbageclunk | fine :) | 01:37 |
hpidcock | wallyworld: I've changed the remotestate/resolver in the uniter for passing through changes in state. https://github.com/juju/juju/pull/11214/ I haven't finished the unit tests, but just a preliminary change to see if you like it. | 01:43 |
wallyworld | ok, ta | 01:43 |
wallyworld | hpidcock: yeah, i think that's ok | 01:56 |
wallyworld | i left a couple of questions | 01:56 |
babbageclunk | anastasiamac: approved the backup format version PR | 01:56 |
anastasiamac | babbageclunk: \o/ | 01:57 |
babbageclunk | kelvinliu: approved the vsphere change too | 01:59 |
kelvinliu | babbageclunk: ty | 01:59 |
anastasiamac | babbageclunk: "plugin" template name copied from elsewhere (I think maybe the juju command)..? want me to give it a diff name? | 02:21 |
anastasiamac | babbageclunk: addressed everything else | 02:22 |
babbageclunk | Just confused me - what about "template"? Or just ""? Not sure. | 02:23 |
anastasiamac | i like "template" :) | 02:23 |
babbageclunk | I guess it would be in the error if there's a syntactic problem with the template, but that's not a problem here since it's hardcoded. | 02:24 |
anastasiamac | yes :D | 02:24 |
babbageclunk | ooh, apparently the name becomes important if you want to have one template that invokes another | 02:30 |
anastasiamac | makes sense | 02:31 |
kelvinliu | wallyworld_: got this pr to add hostNetwork https://github.com/juju/juju/pull/11219 +1 plz | 04:32 |
wallyworld_ | kelvinliu: just did it :-) | 04:32 |
wallyworld_ | ty | 04:32 |
kelvinliu | quick! ty! | 04:33 |
wallyworld_ | hpidcock: i'll try and land this on monday if you can review at some stage whenevr https://github.com/juju/juju/pull/11220 | 06:02 |
hpidcock | wallyworld_: I'll check it out | 06:03 |
nammn_de | anyone with access to spaces want to give review/qa for https://github.com/juju/juju/pull/11183 only contains the patch for the command interface. rick_h wanted to give some feedback on ux, if his vpc is working | 09:18 |
achilleasa | nammn_de: is there any easy way to remove the duplicate "cannot remove space" from the error messages? They look a bit odd; also (minor nitpick) if the reasons are a list maybe we could prefix them with a '-' or something | 09:22 |
nammn_de | Will update the messages accordingly. Makes sense. Will try and update the qa section | 09:24 |
nammn_de | achilleasa: like this? https://paste.ubuntu.com/p/s8kFWfw9MS/ | 09:29 |
nammn_de | hmm I think I can omit the because though | 09:29 |
nammn_de | yeah, I will omit the ... because:\n | 09:30 |
achilleasa | nammn_de: much nicer! Is it also possible to pretty print the constraint list? | 09:30 |
nammn_de | achilleasa: you mean instead of this [{"model-e"} {"application-mediawiki"}] just this "model-e","application-mediawiki" | 09:31 |
achilleasa | nammn_de: or maybe like this: https://paste.ubuntu.com/p/j29kcdpmWZ/ | 09:31 |
achilleasa | I don't remember if we usually quote these or not. Can you check what the other commands do? | 09:32 |
nammn_de | achilleasa: okay, great feedback. Will try to make it look more sexy | 09:32 |
achilleasa | can anyone recommend a lightweight charm that uses storage? | 09:38 |
achilleasa | nvm; swift-storage works for my tests | 09:57 |
manadart | stickupkid: What was the issue with old EC2 accounts/VPCs again? | 11:21 |
manadart | stickupkid: Oh, the flavours are not all available, right? | 11:22 |
stickupkid | manadart, if you don't have one setup (default VPC) - you don't have new instance types | 11:22 |
stickupkid | manadart, like everything is deprecated, but I believe you can use force an instance type, but don't quote me on that | 11:23 |
nammn_de | achilleasa: now it looks pretty | 11:33 |
nammn_de | similiar to your showed output | 11:34 |
achilleasa | nammn_de: nice! | 11:34 |
rick_h | nammn_de: and what is "model-e" there? We don't tend to show the raw tag like that. Should that just be "Found an existing model constraint"? | 12:09 |
nammn_de | rick_h: you are right, I should change that. Will change that for the other tags as well. Like instead of application-mediawiki -> mediawiki | 12:51 |
hml | stickupkid: what do you think? are we ready to move 11093 from WIP status? After updating the form part? | 12:52 |
rick_h | nammn_de: definitely, please look around as to how we would show stuff like that in other places. I think the error messages in the binding calls are done pretty well as good examples | 12:53 |
nammn_de | rick_h: Noted. Will do thanks! | 12:53 |
hml | stickupkid: should have read my email more closely :-) | 13:08 |
nammn_de | rick_h: does this look better? https://paste.ubuntu.com/p/MM4dTt2dkf/ | 13:20 |
rick_h | nammn_de: what does `juju get-constraints mediawiki` show in your example? | 13:22 |
nammn_de | rick_h: i accidentally killed my controller. Need to start it up and create env again | 13:22 |
rick_h | nammn_de: ah sorry | 13:23 |
nammn_de | rick_h: if helps we can quick before daily? Else Im gonna recreate and look later. | 13:23 |
nammn_de | mediawiki is shown as constraint because one can deploy with: juju deploy mediawiki --constraints spaces=bla | 13:23 |
nammn_de | and a constraints entry in the collection is created with the id to that application | 13:23 |
nammn_de | rick_h: https://pastebin.canonical.com/p/PyVfwtY84f/ | 13:30 |
manadart | stickupkid: https://github.com/juju/juju/pull/11221 | 13:58 |
rick_h | nammn_de: but what does "juju get-constraints" show? | 13:59 |
nammn_de | https://paste.ubuntu.com/p/bbnwzYS4ZD/ | 13:59 |
nammn_de | rick_h ^ | 14:00 |
rick_h | y | 14:00 |
rick_h | ty | 14:00 |
rick_h | nammn_de: manadart stickupkid so I'm really thinking we should ignore those for the case of applications. We'll always end up with double output on the constraint and the binding. | 14:00 |
rick_h | what do you think? | 14:01 |
nammn_de | rick_h: afaict I could deploy a application with a constraint but without a binding -> which does not lead to a double output | 14:01 |
rick_h | nammn_de: hmm, ok. Can you test/verify you can deploy with a space constraint but no bindings please? | 14:02 |
nammn_de | that would be my pastebin https://paste.ubuntu.com/p/CnBGZ2C5mX/ | 14:02 |
nammn_de | rick_h: yeah above pastebin | 14:02 |
rick_h | hah ok you beat me to it | 14:03 |
rick_h | ok, that's cool then | 14:04 |
rick_h | nammn_de: so one tweak to the wording on that please. As it reads now it's saying "following existing constraints" but the data afterwards isn't a constraint | 14:04 |
rick_h | nammn_de: so I think it should read more direct like: "db" is used as a constraint on: x | 14:05 |
rick_h | nammn_de: and in the model case it would be: "db" is used as a constraint on: mymodel, mediawiki | 14:07 |
rick_h | how does that read? | 14:07 |
stickupkid | ❯ juju remove-space db | 14:08 |
stickupkid | ERROR cannot remove space "db": | 14:08 |
stickupkid | - Found the following existing constraints: mediawiki | 14:08 |
stickupkid | This is really painful in integration tests | 14:08 |
stickupkid | :( | 14:08 |
achilleasa | stickupkid: nammn_de maybe we should add a --force? | 14:09 |
stickupkid | I wonder if there is a way to do it `--cascade` | 14:09 |
rick_h | achilleasa: nammn_de we should definitely have a --force | 14:10 |
stickupkid | I think using force for this isn't quite right imo | 14:10 |
rick_h | stickupkid: how so? My concern are things like stuck units, machines that didn't delete, etc | 14:10 |
rick_h | stickupkid: basically the machines go unchanged you've just force moved the subnets from the one space name back to alpha | 14:10 |
stickupkid | if you remove-space db, what happens to mediawiki (in that scenario) | 14:10 |
stickupkid | fine, that makes sense | 14:10 |
rick_h | stickupkid: so the subnet moves back to the alpha space | 14:11 |
stickupkid | --force is good | 14:11 |
* stickupkid is happy again | 14:11 | |
rick_h | stickupkid: and now mediawiki is still the same IP address-wise, but the endpoint binding it to alpha | 14:11 |
rick_h | oh ok, /me likes happy stickupkid | 14:11 |
stickupkid | don't want to have that situation around relations, that sucks | 14:11 |
* rick_h cannot process that sentence atm | 14:12 | |
* rick_h sips more coffee... | 14:12 | |
stickupkid | i.e. if you have a relation, you can't remove an application | 14:12 |
rick_h | oh a CMR one? | 14:12 |
stickupkid | or destroy a controller, hence why we brute force it in the integration tests | 14:12 |
stickupkid | yes | 14:12 |
rick_h | if you have a relation you can remove the app, but yes the CMR one is the complicated one that hangs up folks | 14:12 |
rick_h | gotcha, right | 14:13 |
rick_h | definitely don't want more of that | 14:13 |
stickupkid | 100% agree | 14:13 |
nammn_de | will add force. Do we want force to be able to run in any case (controller setting, application binding, constraints)? | 14:13 |
hml | stickupkid: https://bugs.launchpad.net/juju/+bug/1863253 | 14:14 |
mup | Bug #1863253: model cache: publish modelUnitAdd too often with subordinates <juju:Triaged> <https://launchpad.net/bugs/1863253> | 14:14 |
rick_h | nammn_de: yes | 14:14 |
stickupkid | hml, wicked | 14:14 |
hml | stickupkid: anything missing from the description: https://github.com/juju/juju/pull/11093? | 14:18 |
stickupkid | hml, nope, all good from me | 14:19 |
hml | stickupkid: i’ll hit the merge button | 14:20 |
stickupkid | YAY | 14:21 |
stickupkid | hml, me right now https://i.imgur.com/5irRQLe.gif | 14:21 |
hml | :-D | 14:22 |
nammn_de | stickupkid: for better integration test compability, should I remove the "-"? | 14:36 |
stickupkid | nammn_de, from ? | 14:37 |
nammn_de | stickupkid: from the remove-space output you posted above from my example | 14:37 |
stickupkid | nammn_de, nah, that's fine imo | 14:37 |
hml | achilleasa: available to chat on "storing uniter state server-side" ? | 14:49 |
achilleasa | hml: sure. give me 2min | 14:50 |
nammn_de | rick_h: does this output look better? https://paste.ubuntu.com/p/B5yTVs76TV/ | 14:52 |
stickupkid | hml, can you look at my PR for goose, https://github.com/go-goose/goose/pull/77 | 14:56 |
stickupkid | hml, I think that's everything | 14:57 |
hml | stickupkid: sure | 14:57 |
rick_h | nammn_de: looking good, what's the controller settings? Since spaces don't exist until the controller comes up so not sure what that is. I'd also assume it's `controller-config` vs "settings" since that's what the commands are around that | 14:59 |
nammn_de | rick_h: ah, yes controller-config its: juju-ha-space and juju-mgmt-space in case you are running on a controller-model | 15:00 |
rick_h | nammn_de: k, yea let's roll with calling it the same thing throughout (-config) but cool | 15:04 |
nammn_de | rick_h: rgr thanks | 15:04 |
nammn_de | will update with the force option, update qa and link again then | 15:04 |
rick_h | achilleasa: do you have time to help me with my aws vpc setup please? | 15:08 |
rick_h | achilleasa: or not if you're close to school pickup/etc | 15:08 |
achilleasa | rick_h: sure. I am in daily | 15:33 |
rick_h | achilleasa omw | 15:33 |
nammn_de | rick_h: I am trying to fiddle around a good way to use force. Here are the options I see. | 15:42 |
nammn_de | Each option makes the implementation quite different. | 15:42 |
nammn_de | https://paste.ubuntu.com/p/TMGByz6BJk/ | 15:42 |
nammn_de | For the first option i need to use some-kind of dry-run/just run without force and prompt again if error | 15:42 |
rick_h | nammn_de: sec, otp will peek in a minute | 15:44 |
hml | stickupkid: did you get the go test -live working for your goose changs? | 15:44 |
stickupkid | hml, not via the python | 15:45 |
hml | stickupkid: no cli? | 15:45 |
stickupkid | hml, i just switched the default value | 15:47 |
stickupkid | hml, https://github.com/go-goose/goose/blob/v2/neutron/neutron_test.go#L12 | 15:48 |
hml | stickupkid: how are these ports going to be used? after being created? | 15:54 |
stickupkid | hml, ho? | 15:54 |
hml | stickupkid: sure | 15:55 |
rick_h | nammn_de: ok, the first one is good but we have to remove the Y/n so that it's scriptable | 16:01 |
rick_h | if you go force that's it, you're on your own | 16:01 |
nammn_de | rick_h: the Y/n is used in other code as well with the option to provide a -y . | 16:01 |
nammn_de | so I planned to the same | 16:02 |
rick_h | nammn_de: ah ok, then carry on | 16:02 |
nammn_de | rick_h: time for a ho? | 16:06 |
rick_h | nammn_de: sure thing, meet in our 1-1 room? | 16:07 |
nammn_de | rick_h: coming | 16:07 |
achilleasa | rick_h: did the bootstrap work? | 16:45 |
rick_h | achilleasa: no, it failed with a 20m timeout | 16:45 |
achilleasa | :-( did you retry with --debug? | 16:46 |
rick_h | achilleasa: not yet, was otp | 16:47 |
rick_h | and have to run to family lunch. I'll mess with it achilleasa ty for the help! | 16:47 |
* rick_h goes *poof* | 16:48 | |
achilleasa | good luck! | 16:48 |
achilleasa | stickupkid: can you help me track down something? | 17:08 |
stickupkid | achilleasa, maybe | 17:08 |
achilleasa | daily? | 17:08 |
stickupkid | of course | 17:08 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!