[00:33] <tlm> wallyworld: sent you a dm incase your notifications are still playing up
[00:54] <timClicks> is it possible to use `juju deploy` on k8s and require that it's deployed to a pod with a pre-existing CRD? Sort of like --devices but for CRDs?
[00:58] <timClicks> perhaps --constraint would make more sense, if that makes any sense at all
[01:38] <thumper> timClicks: I'm not sure where you're getting with this
[02:19] <tlm> wallyworld: got 5 minutes for HO ?
[02:22] <wallyworld> tlm: i do
[02:23] <wallyworld> hpidcock: if you have any time given there's a meeting tonight https://github.com/juju/juju/pull/11724
[02:28] <hpidcock> wallyworld: sure thing
[03:40] <hpidcock> wallyworld: I'm sorry for the review but here it is https://github.com/juju/juju/pull/11724
[04:11] <tlm> wallyworld: not urgent but when you get 2 minutes can you check I correctly addressed the logging comment in https://github.com/juju/juju/pull/11671
[04:19] <wallyworld> tlm: just finished a call, will look after coffee
[04:56] <timClicks> thumper: you might want to take a look https://discourse.juju.is/t/inline-help-update-juju-deploy/2767/6?u=timclicks
[04:57] <timClicks> it's a fairly aggressive change so I haven't filed a PR
[05:02] <thumper> ok, can we look tomorrow during our call?
[05:02] <thumper> wallyworld: where does the unit cloud container status get updated?
[05:02] <thumper> I can't seem to find it
[05:03] <thumper> hmm... in UpdateUnitOperations
[05:04] <timClicks> thumper: sounds good
[05:12] <wallyworld> thumper: yeah, there
[05:12] <wallyworld> hpidcock: thanks for review, i've made changes but also answered a couple of comments, see if you agree
[05:14] <wallyworld> i gotta go buy dinner, bbiab
[05:14] <thumper> wallyworld: it seems very hard to set to test
[05:14] <thumper> if you have a pointer that'd be helpful
[06:06] <wallyworld> tlm: i think the model controller PR might be missing the loggingConfigUpdater worker
=== arif-ali_ is now known as arif-ali
[08:20] <manadart_> stickupkid or achilleasa: Need a forward merge review: https://github.com/juju/juju/pull/11725. Reasonably substantial.
[08:20] <stickupkid> seems like a lot of deletions
[08:28] <manadart_> stickupkid: From https://github.com/juju/juju/pull/11724.
[08:30] <stickupkid> manadart_, yeah added a comment about him publicly announcing this...
[08:35] <tlm> wallyworld: no it's there
[12:33] <flxfoo> Hi all, quick one
[12:34] <flxfoo> Is there a way to proxying access to instances via the controller? instead of doing direct `juju ssh <id>` ?
[12:35] <achilleasa> flxfoo: did you try 'juju ssh --proxy <id>'?
[12:39] <rick_h> achilleasa:  flxfoo and there's a controller config to set it up to work that way I believe but the user has to have ssh access on the controller for the proxying to work
[12:39] <rick_h> e.g. not great for shared environments/etc
[12:44] <flxfoo> achilleasa: rick_h thanks guys , appreciated...
[14:11] <manadart_> stickupkid: Can you tick a forward merge of the same patch? https://github.com/juju/juju/pull/11727
[14:25] <flxfoo> Hi again
[14:25] <flxfoo> `juju ssh --proxy` just stall
[14:26] <flxfoo> I have a few subnets (class c) to form private subnets... i set spaces related to those subnet...
[14:26] <flxfoo> after deploying, I have a class c IP in "public IP" in juju status
[14:27] <flxfoo> I guess I am missing something here
[14:27] <flxfoo> (so direct ssh or ssh --proxy) don't work
[15:47] <flxfoo> does anybody using aws with private (class c) subnets, instances not having public ips (I mean having private ips) being able to reach them properly? is there a specific settings perhaps?
[15:50] <achilleasa> flxfoo: can you try the script in the "what if I need to ssh into a machine with a private IP" section in https://discourse.juju.is/t/how-to-create-and-use-spaces-on-aws/2115?
[15:54] <flxfoo> achilleasa: will do thanks
[16:13] <stickupkid> hml, CR https://github.com/juju/juju/pull/11728
[16:16] <achilleasa> petevg: got a question. If the plan is to extend the open/close-port tools to work with endpoints, how will charms know whether that feature is available to them (they would get an error running the tool with extra args on older controllers)
[16:17] <achilleasa> do we want to provide a new open/close-port-for-endpoints tool which charms can discover on their path?
[16:19] <achilleasa> stickupkid: have we done anything similar in the past? ^^^
[16:20] <stickupkid> achilleasa, not a clue on that one
=== arif-ali_ is now known as arif-ali
[18:15] <rick_h> hml:  ping, do you know if there's one juju operator pod per unit in a k8s model or one pod per application?
[18:16] <hml> rick_h: i believe it’s per application.
[18:28] <petevg> achilleasa: whoops. Never followed up on your question. Sorry. Short answer: I don't know. Long answer: we should probably put together a list of possibilities, and pick the least bad.
[19:54] <flxfoo> achilleasa:thanks for your answer... I think I got more detail though
[19:54] <flxfoo> the controller is on a different vpc
[19:55] <flxfoo> we enabl peering between the two vpc
[19:55] <flxfoo> then now if I ssh an instance (with the ssh key) from within the controller I can connect.
[19:56] <flxfoo> but the client does not
[20:08] <flxfoo> achilleasa:the script endup with usage: nc [-46CDdFhklNnrStUuvZz] [-I length] [-i interval] [-M ttl]
[20:19] <flxfoo>  achilleasa using ssh -J seems to be working
[20:27] <flxfoo> achilleasa:I use your script to base another one with a one liner... should I post it on the page you mentioned?
[22:17] <tlm> thumper: got 5 minutes for HO ?
[22:47] <thumper> tlm: sure