=== salmankhan1 is now known as salmankhan [09:57] timClicks, are you around today? are we still doing the charming doc meeting? I know both Facundo and chipaca are out, and I didn't know if you were or not [09:59] jam: yes, I had intended on being there.. but I'm happy to chat you to later in the week [09:59] timClicks, I'm happy to be there, we can talk through what you commented on last week. [10:00] perfect [10:00] jam: would you like to chat now rather than in 30 mins? [10:01] jam: actually, I've just noticed that niemeyer has indicated that he would like to take part === salmankhan1 is now known as salmankhan [14:02] hml: can I get a CR on https://github.com/juju/juju/pull/11526 please? QA steps will take a while... sorry ;-) [14:08] achilleasa: sure [14:42] achilleasa: line 46 of: https://pastebin.canonical.com/p/b46yRMvvmJ/ [14:42] ran the k8s charm qa steps. unsure of the last result [14:45] hml: that looks correct. If you don't set the application databag via mariadb-k8s/0 it shows up empty [14:46] note that these charms are pre 2.7 so they are storing the credentials in the unit databags [14:46] instead of the app one [14:46] achilleasa: am i using the correct charms/ bundle then? [14:47] yes. you need to run the 'relation-set' command from the QA steps to populate the databag [14:48] if you run the last command on rc0 you would hit the bug [14:48] achilleasa: i only ran relation-set on 1 of the units, not both [14:49] hml: yes, you ran it on /1 which is not the leader so nothing got set [14:49] (the set reported an error in the pastebin output) [14:51] achilleasa: crap - hit a new bug after i added a model to the controller to start over…. : ERROR cannot deploy bundle: cannot deploy application "mariadb-k8s": cannot add application "mariadb-k8s": series "kubernetes" in a non container model not valid. [14:52] achilleasa: nm… let me figure this out. s omething is off in my config === negronjl is now known as negronjl_away [15:38] achilleasa, https://github.com/juju/juju/pull/11527 adds appropriate behavior to 'juju debug-code' [15:38] and adds tests of how debugAt interacts with shell scripts. [15:39] * jam eod === negronjl_away is now known as negronjl === negronjl is now known as negronjl_away [16:01] achilleasa: is it necessary to update the uniter api version? i’m having an internal debate === negronjl_away is now known as negronjl === negronjl is now known as negronjl_away [16:09] hml: since rc0 is out we need to if we want to ensure a proper upgrade path, right? === negronjl_away is now known as negronjl === negronjl is now known as negronjl_away [16:24] achilleasa: rc0 would be fuzzy for me with this. better safe than sorry though [16:30] achilleasa, we have only officially released beta1, right? [16:30] rc1 is the 'edge' not officially released [16:30] achilleasa, so we have to have a way to *upgrade* from beta1 to final, but that doesn't mean beta1 has to be compatible with final [16:31] achilleasa, so if what you are changing would affect *upgrade* then you definitely need to version it. if there is an error, but upgrading still works, that is ok. [16:35] jam: I bumped it to be on the safe side as I was not entirely sure about facade schemas (my change adds a new call) and pylibjuju [16:46] achilleasa: a comment and a question. === negronjl_away is now known as negronjl === negronjl is now known as negronjl_away === negronjl_away is now known as negronjl === narindergupta is now known as narinderguptamac === _thumper_ is now known as thumper [21:05] * thumper is looking at a very mocked worker trying to work out how to write a test for the current failure [21:26] :( I can't help feel that these tests are racy [21:26] perhaps I don't understand the mocks [22:43] kelvinliu: if you're around could you take a look at https://github.com/juju/juju/pull/11528 please? [23:02] babbageclunk: sure [23:03] babbageclunk: could u take look this one as well https://github.com/juju/juju/pull/11505 please?