=== ses is now known as Guest47761 [00:34] Bug #1556514 opened: Juju bootstrap does not assign and use floating IP on openstack [00:40] Bug #1556514 changed: Juju bootstrap does not assign and use floating IP on openstack [00:49] Bug #1556514 opened: Juju bootstrap does not assign and use floating IP on openstack [04:28] Bug #1532841 changed: cannot add charm to storage [04:35] Bug #1532841 opened: cannot add charm to storage [04:38] Bug #1532841 changed: cannot add charm to storage [15:54] Bug #1556630 opened: Timeout github.com/juju/juju/cmd/jujud/reboot [16:15] Bug #1555291 changed: TestStartInstancePopulatesNetworkInfoWithoutAddressAllocation fails on centos [20:46] thumper: https://github.com/juju/juju/pull/4706 [20:46] two things [20:46] 1, do you have lxc installed ? [20:47] i don't and the tests always pass reliably [20:47] but the test was ignoring basically all the error handling [20:47] so I added it back in [20:47] at least the test won't hang now [20:47] yes I have lxc installed [20:48] thumper: PTAL https://github.com/juju/juju/pull/4706 [20:48] it reduced the test run time from 10s of seconds to milliseconds [20:49] i cannot get it to fail on my machine [20:49] ask you have the failing sample, can ypu try please [20:49] I'll try [21:05] tests pass here now [21:05] davecheney: but I'm not sure why [21:05] how does your branch fix it? === _thumper_ is now known as thumper [22:50] thumper: i don't believe that this change will make the test pass where they previously failed [22:50] but previously the test didn't fail, it hung [22:50] with this change [22:50] you'll get an immediate error [22:50] which is better than the test hanging [22:50] then the test runner being killed and we get zero output [22:50] right [22:51] why did they hang? [22:51] it isn't clear to me [22:52] no idea [22:52] i cannot reproduce the problem on my machine [22:52] and when we do it in CI [22:52] the test runner kills the job before we get any output [22:55] well, even with your stress.sh it passes every time ow [22:55] now [22:56] thumper: if you can make it hang on your machine without that CL [22:56] can you run the test with debug turned up to 11 pls [22:56] * thumper tries [22:57] hmm... [22:57] without your changes... [22:57] it hangs first time [22:59] what output do yo get [23:00] are any of those log lines firing ? [23:00] * thumper looks [23:00] * thumper added -check.vv [23:01] [LOG] 0:15.878 WARNING juju.cmd.jujud.reboot Waiting for containers to shutdown: [lxd:juju-c0518c0b-97b5-47ab-8c1b-541906fb2c73-machine-0] [23:01] juju-c0518c0b-97b5-47ab-8c1b-541906fb2c73-machine-0 | STOPPED | | | PERSISTENT | 0 | [23:02] hmm, lxc-ls is being *ahem* patched [23:02] you should not see that [23:02] this is lxd [23:02] not lxc [23:02] lxc list [23:02] we'd that would explain it [23:02] welp [23:03] why did it pass with your patch then? [23:03] * thumper is confused [23:03] i don't know [23:04] i cannot get it to fail on my machine [23:04] got any lxd containes? [23:04] i have to do this surgery via remote [23:04] nope, don't have lxd installed [23:04] ah... [23:04] that'll be why it is fine on your machine [23:04] because lxd isn't initialized [23:04] that's some great test isolation we have [23:04] fricken amazing [23:05] * thumper heads to lunch [23:16] cherylj: I don't have permission to close this bug https://github.com/juju/juju/issues/4705#issuecomment-195956659 [23:16] may I please have that permission