| [21:41] <thumper> morning folks | |
| === wallyworld_ is now known as wallyworld | |
| [22:45] <thumper> hi waigani | |
| [22:45] <thumper> wallyworld: morning | |
| [22:45] <wallyworld> hey | |
| [22:45] <waigani> thumper: hello | |
| [22:45] <thumper> wallyworld: https://bugs.launchpad.net/juju-core/+bug/1360286 | |
| [22:45] <mup> Bug #1360286: ppc64el and i386 unittests cannot find tools <ci> <regression> <juju-core:Triaged> <https://launchpad.net/bugs/1360286> | |
| [22:45] <thumper> wallyworld: my guess is axw's bootstrap refactoring | |
| [22:45] <thumper> wallyworld: do you think we should revert or fix? | |
| [22:46] <thumper> wallyworld: blocking CI | |
| [22:46] <wallyworld> could be, i'll look if my network connection satys up long enough :-( | |
| [22:46] <wallyworld> let me look first | |
| [22:46] <waigani> thumper: checking the user exists I can only see an unexported func state.checkUserExists | |
| [22:47] <thumper> waigani: what is this for? | |
| [22:47] <waigani> thumper: AddEnvironmentUser | |
| [22:47] <waigani> thumper: checking if user exists before adding it | |
| [22:47] <thumper> waigani: ok, there are two main possibilities | |
| [22:48] <thumper> waigani: 1) check at the apiserver level using state.User(name) to load the user | |
| [22:48] <thumper> 2) look inside the actual state method to add an environment user | |
| [22:48] <thumper> I'd probably prefer 2. | |
| [22:48] <waigani> thumper: okay, also how do we tell if they are local? | |
| [22:49] <thumper> waigani: the tag Provider method is "local" | |
| [22:49] <waigani> thumper: thanks | |
| [22:54] <wallyworld> thumper: the failing test can be removed since the logic has changed - i'll do a branch with it skipped to unblock ci and confirm with andrew and then remove it | |
| [22:55] <thumper> wallyworld: cool | |
| [23:15] <thumper> wallyworld: http://juju-ci.vapour.ws:8080/job/github-merge-juju/buildTimeTrend | |
| [23:15] <thumper> wallyworld: FYI looks like Build #374 (21-Aug-2014 07:52:39) axw environ-bootstrap-upload-tools is the one that started big time changes | |
| [23:22] <wallyworld> thumper: our build times have also been going up because tests are really unreliable and we now fail back to -p 2 most times | |
| [23:22] <thumper> ah | |
| [23:23] * thumper suggests that we only check once | |
| [23:23] <thumper> fail faster if we are going to fail | |
| [23:27] <wallyworld> trouble is the 2nd run normally works | |
| [23:27] <wallyworld> but is much slower than with full parallelism | |
| [23:28] <wallyworld> we just need to fix the freaking tests | |
| [23:36] <thumper> wallyworld: agreed | |
| [23:37] <thumper> the replicaset test seems to be overly fragile | |
| [23:37] <thumper> and long | |
| [23:37] <thumper> can we not force that out of our normal test run into something more suitable? | |
| [23:37] <thumper> and write a robust test? | |
| [23:37] <wallyworld> you would think so | |
| [23:38] <wallyworld> also, jujud tests are horrible | |
| [23:38] <wallyworld> and there's a failing uniter test | |