[05:20] wallyworld: btw, PTAL https://github.com/juju/juju/pull/10586 :) [05:21] wallyworld: i've requested u as reviewer on github when I proposed but obviously it's way too subtle :) [05:21] ok [05:21] didn't see it [05:21] all good [05:21] most gh email is filtered as it's too noisy [05:23] :) [05:34] anastasiamac: see comment/question about the txn logic [05:34] wallyworld: k [05:35] wallyworld: i'd rather keep docxists assert but will add buildttxn loop and the test [05:36] in that case, also add assert for $ne Dead [05:36] wallyworld: k [05:36] to match initial query [05:36] wallyworld: funny but this code was almost verbatum from upgrade permissions... i wonder then if there'd b a problem there too.. [05:37] upgrade steps you are talking about" upgrades don't need to deal with concurrent access to db [05:37] s/"/? [05:55] altho mayb just rmoving assert is fine... u r right, we don't care if we cannot clear cred on a gone model.. [06:46] i think in this case it may be ok to not care so much [06:48] wallyworld: meh.. alreay pushed tansactionality and buildtxn ... want to look? [07:13] Can I get a tick on a forward-merge? https://github.com/juju/juju/pull/10590 [07:14] jam: Ta. [07:30] anastasiamac: sure, looking now [07:31] anastasiamac: still lgtm [07:33] wallyworld: \o/ [07:33] ah manadart if only that merge was a couple of hours late :) [07:34] i'll just cherry-pcik my change into develop tomorrow :D [07:34] anastasiamac: If the merge fails, I'll add yours :) [07:36] manadart: mine just went for landing now... [07:39] into 2.6 that is.. [08:51] wallyworld: and the cherry-pick for develop PTAL https://github.com/juju/juju/pull/10591 [08:52] manadart: mayb u could a +1? srsly just cherry picking :) [08:55] anastasiamac: Tick! [08:55] manadart: \o/ [10:31] achilleasa, any chance of a CR, this is the changes I talked about last week https://github.com/juju/bundlechanges/pull/57 [10:31] achilleasa, fixing the tests took a lot longer than i hoped :\ [10:32] stickupkid: sure thing. looking in a few min [10:33] achilleasa, it's a shame GUIArgs has to live on, as it's not a very efficient way to transport information... i.e. you need empty values for all options, even if they're empty [10:36] stickupkid: that's only being used for the GUI bits right? We could kindly ask them to switch to the map version instead ;-) [10:36] achilleasa, yes, i agree [11:29] stickupkid: approved. Yeah, I remember updating the tests to include the empty channel in the GUIArgs output was a pain too [13:31] anybody experienced "twd4f3: Unable to allocate static IP due to address exhaustion", and have a fix (not sure what network is exhausted, and from the maas there are plenty of IPs in all networks)... [13:32] (ignore twd4f3 its just the MAAS id) [14:25] manadart: can or should 1 subnet be in 2 spaces? [14:25] hml: No. [14:26] manadart: that’s what i thought. juju may be broken there. stumbled on it while updating RelationUnitSuite.TestNetworksForRelationWithSpaces [14:27] for space-3 [14:27] hoping was a copy paste error in the test setup [14:30] look like we never check [14:49] found the check… up in the apiserver code. but that doesn’t prevent state tests from doing interesting things. === salmankhan1 is now known as salmankhan [15:00] rick_h, i've updated to pylibjuju PR, I want to now add more unit tests around the type classes, as it should be much easier to do so. Can you look to see if you've got any concerns [15:00] rick_h, https://github.com/juju/python-libjuju/pull/350 [15:00] stickupkid: rgr, otp but will look in a few [15:12] stickupkid: feedback in, looks much sweeter! [15:13] rick_h, the nice thing is, we can use mocks to test the run methods now! [15:13] stickupkid: exactly [15:13] rick_h, yarp, just adding some tests [15:13] stickupkid: sorry to create the work but I think it's on a nicer path for sure [15:13] rick_h, i'd rather have it right, 100% agree [19:27] Hi all! [19:28] has anyone used openstack as a cloud provider in Juju? [19:31] I'm getting this error: ERROR juju.cmd.juju.commands bootstrap.go:697 failed to bootstrap model: no image metadata found [19:32] also: skipping index "http://my_ip:8080/simplestreams/images/streams/v1/index.json" because of missing information: index file has no data for cloud {Louros https://mykeystoneurl:5000/v3} not found [19:36] same steps have been working about a year ago.. [19:36] yan0s: check the index.json to see if the exact url is given: https://mykeystoneurl:5000/v3 [19:37] yan0s: i’ve seen where it’s https://mykeystoneurl:5000/v3 vs https://xx.xx.xx.xx:5000/v3 where technically they are the same, but it fails. [19:37] it’s a straight string match, no lookup [20:07] hml: it is the same in index.json [20:09] yan0s: which version of juju? [20:09] yan0s: any possibility of a link to or pastebin of the the index.json file? and the full bootstrap —debug output please? [20:10] 2.6.8-bionic-amd64 [20:11] https://pastebin.com/mrfzRMi3 [20:13] actually here is my simplestream uploaded http://62.217.82.41:8080/simplestreams/images/streams/v1/ [20:17] yan0s: there was a mistype in creating the simplestream files… “https:///keysto” 3 slashes instead of 2. [20:18] oh god.. [20:19] I'm so sorry for wasting your time [20:19] yan0s: no problem. another set of eyes can be a good thing [20:19] :-) [20:20] yan0s: great to hear that this was an easy fix :) [20:23] yan0s: do let us know if you have any further problems, we will be delighted to help [23:18] https://discourse.jujucharms.com/t/square-brackets-in-env-vars/2024 if anyone knows what goes on under the hood and what I have to escape to get this working, that would be very nice indeed! :) [23:43] some more intermittent test failure fixes https://github.com/juju/juju/pull/10594 [23:44] wallyworld: you may have commentary for magicaltrout on that post [23:45] magicaltrout: I'm guessing that it is being parsed as a yaml array [23:45] * thumper thinks [23:46] magicaltrout: now I read it a bit more, it does seem very strange [23:46] magicaltrout: you can use yaml type forcing [23:46] I'd expect [23:46] magicaltrout: also, what is dc4? [23:47] magicaltrout: you also don't mention where you are putting it in juju [23:47] wallyworld: I think I know what is going on though [23:48] wallyworld: I bet if you take a string like '["hello"]' and parse it as yaml, then write it out as yaml you get ["hello"] [23:48] thumper: sorry, was otp [23:48] which the second parser deeper is treating as a list