[14:35] Laney, i am suspicious about ros-ros-comm adt test on cosmic on i386 triggered by python2.7/2.7.15-4ubuntu4 -> has it been continiously running for multiple hours now; killing the VM to oblivion, and autopkgtest cloud continiously auto-restarting it, or some such? [14:35] Laney, it should only take just over an hour to complete, but it seems to be actually running and running for hours, and I think I did catch a running log with a much higher command, than the current command9 before. [14:37] how many should it have? [14:37] try running it locally? [14:38] i think command24 is the highest. [14:38] also no idea why i started talking in this channel [14:38] good for it to have some activity :-) [14:39] it could be that it times out [14:39] and we need to put it in long_tests or something [14:59] Laney, but like i see no long / test runs about it from the results page... [14:59] Laney, are there any more logs that you might have access about it? It does look suspicious. [14:59] i.e. i'd would want to know if it was autorestarted or not. [14:59] http://autopkgtest.ubuntu.com/running#pkg-ros-ros-comm [15:07] it finished?!?!?! [15:10] Laney, well yeah, weird. now waiting 5-10 minutes for the results to appear in all the right pages. [15:10] Laney, must be scared of you logging-in ;-) [15:10] heh [15:13] Laney, it's back from the dead. running "still" now at 2h.... http://autopkgtest.ubuntu.com/running#pkg-ros-ros-comm [15:13] Laney, did it restart? [15:14] it's done command1 and now doing command2, when i pinged you it was at command9 at least. [15:14] so it did jump back. [15:14] i fear it might be doing this in perpetuity. [15:19] ok lemme save the log [15:19] catch it in the act of being naughty [15:34] Timed out waiting for ssh. Aborting! Console log: [15:34] ------- nova console-log b6f59c2d-bad8-46fe-9399-9ca91765b66c (adt-cosmic-i386-ros-ros-comm-20181003-131224) ------ [15:34] --------------------------------------------------- [15:35] ------- nova console-log af7053b5-66b5-4396-a600-a3b09634f399 (adt-cosmic-i386-ros-ros-comm-20181003-131224) ------ [15:35] ERROR (CommandError): No server with a name or ID of 'af7053b5-66b5-4396-a600-a3b09634f399' exists. [15:35] --------------------------------------------------- [15:35] ERROR (CommandError): Unable to delete the specified server(s). [15:37] dunno what that even means, the server failed to provision?????? [15:44] muahahhaha [15:46] Laney, meanwhile, it seems to me, it is written like that, because they did not want to write a shell script. [15:46] Laney, i think all of this can run on a single VM.... instead of gazzilion VMs. [15:47] it does mean they get separate results which can be nice [15:48] but maybe this is taking it a bit far [15:58] if separates logs is important, i can create separate logs in artifacts per test-command..... [15:58] 144 VMs vs 6 VMs is imho a change worth doing here [15:59] given no deps are conflicting, and there is no breaks-testbed, and there is no needs-root, etc [15:59] this is abuse of resources [16:01] in the future we'll have the ability to skip / track regressions per test [16:01] in that sense splitting your tests up in to separate autopkgtests is sensible [16:02] but I wouldn't object if you were to consolidate this a bit :-) [16:02] well, ideally we would want some kind of a reporting api thing, such that a single test can export a report of multiple tests. [16:02] it'll probably eventually finish by luck [16:02] this runs trivial single g++ on a oneline file, testing api compilation..... [16:03] yes OK, that probably doesn't need to be N different tests ... [16:03] * Laney didn't look at what they actually are [16:03] sure. [16:03] if not these failures, i wouldn't either =) [16:05] in total it is a 35s test; with hours of VM setups =) [16:08] ;_; [16:09] CLOUD! [16:44] it got up to test 11 this time and I'm about to have to go :( [16:44] 12* [16:45] stupid thing