[06:44] <akaradag> hello again, does anyone know ubuntu's automated test method for os installation
[07:00] <jpds_> akaradag: Yes, what about it?
[07:01] <akaradag> good, i am searching a method to test os installation as automatically
[07:02] <akaradag> what is ubuntu's method for that?
[07:02] <jpds_> preseed.
[07:03] <jpds_> Same as Debian.
[07:29] <jibel> akaradag, https://help.ubuntu.com/14.04/installation-guide/i386/apb.html
[07:30] <dkessel> good morning
[11:19] <elfy> yay - can install to whole disk again \o/
[17:42] <average_guy> I am chasing a bug I filed #1352905 and have found the solution - dunno wht to do nxt
[17:43] <average_guy> testdrive pulls the iso listed on the manifest, the tracker where is says "Link to download information" does not show the correct link for zsync
[17:44] <average_guy> hence http://iso.qa.ubuntu.com/qatracker/milestones/315/builds/74832/testcases/1337/results
[18:06] <average_guy> jibel or stgraber  ^^  I am told you guys are the ones I should mention this to
[18:10] <balloons> bug 1352905
[18:11] <balloons> average_guy, you can check the image testdrive pulls vs the one you downloaded
[18:11] <balloons> look in ~/.cache/testdrive/iso to find the images
[18:12] <balloons> compare it with what you downloaded.. grab the md5sums and check both
[18:12] <balloons> make sense?
[18:12] <average_guy> i peeked under the hood of testdrive to see how it gets it's img balloons
[18:12] <average_guy> it pulls from the manifest
[18:12] <average_guy> the zsync link does not match
[18:12] <balloons> sure, but the bug is a little confusing.. or I'm confused
[18:13] <balloons> are the images different?
[18:13] <ianorlin> yes but is it same image as it could be a mirror?
[18:13] <balloons> well check the md5's and see
[18:20] <average_guy> they do not match
[18:21] <balloons> i replied to the bug with some questions.. if you can provide the details it should help pinpoint the problem
[18:21] <average_guy> ok
[18:21] <balloons> thanks ;-) and well spotted
[18:22] <average_guy> ty :) am pretty proud of myself. It has been "bugging" me
[18:24] <elopio> balloons: hello.
[18:25] <elopio> balloons: your branch with the fixes for the warning is great, thanks for that.
[18:25] <elopio> balloons: but your test is still no good, because I can't make it fail.
[18:25] <elopio> and I can't find a way to make it work. I've tried all the ways to reload a module I have found.
[18:26] <elopio> balloons: as it's unlikely to regress because we will never touch the emulators module again, I'm ok with removing the test, if you agree that's a good idea.
[18:42] <balloons> elopio, you don't think importing in the testcase works?
[18:42] <balloons> elopio, I am ok with removing it.. it's probably a meaningless test
[18:43] <elopio> balloons: it doesn't work because by the time you hit it, it's already in sys.modules
[18:43] <elopio> so the second time you do the import, the warning is not logged anyway. So the test will always pass.
[18:44] <elopio> it was a good idea to test it, but now we will spend more time trying to get it right than the value we will get out of it.
[18:44] <balloons> elopio, ok fair enough.. I'll just drop it
[18:44] <elopio> I tested it manually, and all your changes solve the issue.
[18:48] <balloons> I removed and pushed
[18:58] <balloons> average_guy, so the tracker it is then?
[18:59] <average_guy> yes balloons, the tracker shows the wrong link for zsync'ing
[18:59] <balloons> average_guy, mmm..
[18:59] <balloons> i think I see the problem actually: http://cdimage.ubuntu.com/ubuntu-server/daily/current/
[18:59] <balloons> the manifests are old
[19:00] <average_guy> I didn't even know that cld happen :(
[19:00] <balloons> it's just a bit odd
[19:00] <balloons> I'm confused as well
[19:01] <average_guy> I just know I get 2 diff test results, and dont wanna test everything twice
[19:03] <balloons> average_guy, so testdrive is indeed correct here. This is sort of a weird case
[19:04] <balloons> my guess is the issue is images are not passing the automatic testing and thus the "current" image is out of date from the daily built image which the tracker links to
[19:04] <balloons> heh, though actually that means testdrive is at fault
[19:05] <average_guy> I see
[19:05] <balloons> the trouble is, do we want a result for an image that failed automatic testing; the answer is no, that's why we gate them
[19:05] <balloons> it's just never gotten so out of sync before
[19:05] <balloons> so the issue has never really presented itself
[19:05] <ianorlin> it does say the image is oversized on the testing page
[19:06] <balloons> so technically, on the tracker, you should grab the daily image as listed on the tracker.
[19:10] <average_guy> gotcha balloons, trust the tracker - the manifest is the problem
[19:11] <balloons> yes trsut the tracker.. the problem is that testdrive only grabs known good images.. which is technically what we want you to test
[19:11] <balloons> it's worthy of discussion on the mailing list
[19:11] <balloons> it's an interesting problem to have it carry on and be out of sync