[01:07] babbageclunk: a few questions on the metrics thing [01:08] thumper: thanks, looking [01:58] thumper: take another look? [02:03] babbageclunk: ack [02:03] ta [02:12] I have to say, this new model summary watcher driven by the cache is turning out to be real smooth [02:26] thumper: the main reason I wasn't going to include timings for errors was that the worker bounces when it gets an error, so they're kind of outside the normal request handling. Unlike in the store, where errors are reported back to the client, so errors are part of the usual workload. Might be wrong though - want me to add them in? [02:27] ah... so if we get one, it has a very small likelihood of every being observed in the metrics? [02:27] yup [02:27] ok, I'm fine with us not having that then [02:27] we have a worker start count [02:27] which gives an indication [02:28] hmm... [02:28] * thumper thinks about getting worker starts as an engine metric [02:28] ha [02:28] * babbageclunk hits merge [02:34] easy test review for someone? https://github.com/juju/juju/pull/11108 [02:34] * thumper looks [02:36] thanks [02:37] babbageclunk: lgtm [02:43] thanks again! [05:09] kelvinliu: haven't quite finished reviewing your pr yet, sorry! it's looking great though - will finish it tomorrow morning.. [05:10] babbageclunk: no rush, take ur time, thanks! [07:49] Hi [11:10] manadart achilleasa: I keep getting it wrong with the facades, could one help me out and take a quick look in regards to the facade bumpings? Seems to be correct to me. The code itself still needs to be tested, but cannot seem to get the correct facade versions https://github.com/juju/juju/pull/11088 [11:11] nammn_de: Looking. [11:19] manadart: ahhh sorry it seems to work now! [11:20] manadart: looks like I was having too many controller running at once and accidentally upgraded the wrong one and got confused, sorry again! [11:21] nammn_de: I was going to say, it looks right. You also need to mask the new method for the old facades, so something like `func (api *APIv5) ShowSpace(_, _ struct{}) {}` [11:21] manadart: ahh great soo looking still helped me :D [13:10] The formatters used by jujuc cmds seem to append a line-feed. Anyone knows whether I can somehow suppress this? [13:15] jam: This confuses my storage implementation as it assumes that data is available and tries to unmarshal it. I can probably trim the last LF (payload may contain linefeeds) as a workaround but it's an ugly hack (plus I would need to detect the os LF char) [13:22] achilleasa: there has been a fair bit of discussion around line endings. in bash, it auto-trims the (foo=$(relation-get bar) just returns the contents), and it is a lot friendlier for human consumption [13:22] achilleasa: I think you're hitting this because you're doing pure bytes instead of str [13:22] achilleasa: ideally, we wouldn't use pickle for this [13:22] IMO [13:23] jam: for now I will just call '.rstrip()' assuming that the payload is encoded and not just plain-text [14:45] achilleasa: hml: so we talked for a while on the 'should process-hook always be called', I think we're back at "yes", and I'll talk through the details we talked about a bit later. (my EOD) [14:46] jam: I can be available a 1/2 before standup tomorrow to chat? [14:46] 1/2 hour === Bandu is now known as sevillafc [20:28] Hi. Any chance I can get some eyes on my promulgation request for the wordpress charm? https://discourse.jujucharms.com/t/promulgation-request-for-wordpress-charm/2494 [21:14] cory_fu: do you know who is currently handling promulgation requests? [21:15] thumper: Anyone in the charmers group can do it with `charm promulgate` [21:15] * thumper goes to see who is in the charmers group [21:15] thumper: But if there's a pending request, I'm happy to do it [21:16] Oh, I see there is. One sec [21:16] cory_fu: thanks [21:19] thumper, mruffell: Done, but there was one hiccup. The old charm was trusty only, and the new one is xenial and bionic. So now trying to deploy on trusty with either cs:wordpress or cs:trusty/wordpress will fail, and it also reset the rev number down to 0 [21:20] I don't know how much of an issue that will be, and the old rev was only 5, so it wouldn't be hard to fast-forward past that. [21:21] it is weird that the rev number dropped to 0 [21:21] cory_fu: I couldn't support trusty, as wordpress has moved past the php version in trusty's repos [21:21] this charm is all good for xenial, bionic, and focal when it comes out [21:21] thumper: Because it's technically a different name (cs:wordpress vs cs:trusty/wordpress) [21:22] mruffell: +1, I doubt anyone is going to be doing any new deployments onto trusty these days anyway [21:22] cory_fu: ah... ok [21:24] anyway, thank you very much thumper and cory_fu. I will let customer know that wordpress is good again. Plus it makes the documentation where wordpress is used as an example make sense too. [22:25] thumper: Are you able to invite other users to that openstack tenant? Management > Project Users > (there should be an Invite User button). If not, I might have to add some role. [22:28] evhan: I'm not sure I'm seeing the same thing you are [22:28] the Projects element in the Identity section gives me a 403 [22:29] Hrn, that'll be related, 1s caller... [22:29] thumper: Perhaps now? [22:30] stil 403 [22:32] Hm, and I don't suppose if you navigate directly to https://dashboard.cloud.catalyst.net.nz/management/project_users/ it'll work? [22:33] different type of 403, rather than top level 403, which I get for /identity, I get the page rendered with header and side bar, but contents says "You are not authorized to access this page Login" [22:35] Curse you, web stuff. [22:35] Alright, I'll have to check about that and get back to you, soz. [22:55] review for a backport of my test tweak to 2.7 (now that 2.7.1 is approved)? https://github.com/juju/juju/pull/11115 [22:59] * anastasiamac looking [22:59] babbageclunk: +1 [23:00] thanks! [23:11] hi