[03:11] <hpidcock> wallyworld: was there any more information on these py libjuju tasks? do you remember if any bugs were filed?
[03:13] <wallyworld> hpidcock: pretty sure there's a bug for the storage one, i'll see if i can find it
[03:13] <hpidcock> danke
[03:14] <wallyworld> hpidcock: https://github.com/juju/python-libjuju/issues/361
[03:15] <wallyworld> i can talk through the issue if you need clarification
[03:17] <hpidcock> no that helps a lot thanks
[03:19] <wallyworld> hpidcock: with the constraints bug, it's just a case of the api in the libjuju client not being implemented
[03:19] <hpidcock> wallyworld: yeah I think I should be able to figure this one out, I'll probably hit you up if I have any questions
[13:24] <manadart> Minor acceptance test fix: https://github.com/juju/juju/pull/10974
[14:58] <manadart> achilleasa: Quick audit. The only higher order usage of device/address provider ID is AllProviderInterfaceInfos, which itself is unutilised.
[14:58] <manadart> Except in tests.
[15:09] <nammn_de1> manadart: do you have some time to talk about this https://github.com/juju/juju/pull/10967 and which edge cases you see in generall? I added mine to the the pr description
[15:14] <nammn_de1> stickupkid: i added those 'edge'-cases i had in mind into the description. If you have some in mind feel free to add/comment them as well. I don't have that many in mind right now
[15:15] <stickupkid> nammn_de1, will do
[15:15] <nammn_de1> because right now I have 2 and the code only catches one, But there are probably more which i just cannot see
[15:17] <manadart> nammn_de1: Give me a few.
[15:37] <stickupkid> CR anyone https://github.com/juju/juju/pull/10975
[15:47] <nammn_de1> manadart: sure just ping me if you have time
[15:56] <achilleasa> manadart: so the solution we discussed shouldn't impact anything then, right?
[15:56] <manadart> achilleasa: No current ops. We're OK for now.
[16:04] <manadart> nammn_de1: Do you want to HO?
[16:04] <nammn_de1> manadart: sure!
[16:07] <manadart> nammn_de1: I'm in Daily.
[16:34] <stickupkid> this is interesting - get a failure attempting to output juju status tabular with branches https://github.com/juju/juju/runs/329443547
[16:37] <nammn_de1> anyone got 10 min time to compare vpn settings again? :D I just got out of a call with the support and we did not come to a conclusion :D
[16:38] <stickupkid> nammn_de1, yap - 2 secs
[16:39] <nammn_de1> while at it, does anyone know of any canonical service which can only be reached with vpn, excluding our own jenkins stuff (10.125.0.203)
[16:41] <stickupkid> nammn_de1, in daily
[17:18] <hml> stickupkid: looking
[17:42] <nammn_de1> dammn i keep forgetting that i should change the acceptancetest for 2.7  first and merge on develop..
[17:42] <nammn_de1> Here is the backport for 2.7 https://github.com/juju/juju/pull/10976
[17:42] <nammn_de1> manadart ^
[17:55] <hml> stickupkid: approved
[23:05] <timClicks> what do providers need to do to support multiple instance types? We allow openstack users to provide their own image metadata and have received requests to allow vSphere users to do the same..
[23:09] <babbageclunk> timClicks: is instance type what you mean here? I think instance types are less flexible than picking CPUs/ram/disk directly, aren't they?
[23:10] <timClicks> am not 100% sure. if I were to provide my own OVF, for example.. would I register that as an instance type?
[23:12] <babbageclunk> just looking at what's in the ovf...
[23:13] <timClicks> the end game would be to able to support centos7 images, for example
[23:13] <babbageclunk> Right, I think that's definitely image rather than instance type
[23:14] <babbageclunk> It doesn't seem like instances are constrained to only have the cpus/ram/disk specified in the ovf
[23:15] <babbageclunk> but there's not an obvious way to use a different disk image.
[23:15] <timClicks> no, because we customise that base template later
[23:17] <babbageclunk> the vmdk is pulled from cloud-images.ubuntu.com - I guess if we could swap that out, it should allow substituting the image
[23:19] <babbageclunk> We do have something called CustomImageMetadata in instance config - I'm trying to trace that down
[23:36] <babbageclunk> timClicks: It seems like we *might* be able to make a substitute image metadata dir using `juju-metadata generate-image` that contains a different ova file? But I haven't played around with it
[23:38] <babbageclunk> I'm not at all clear how to run it
[23:39] <timClicks> btw I have updated the default theme in discourse
[23:39] <timClicks> the theme is now more... on theme