mozilla :: #automation

19 Apr 2017
07:05whimbooFlorinMezei: do we want to check bug 1303834 together?
07:05firebothttps://bugzil.la/1303834 NEW, nobody@mozilla.org [tier-3] Intermittent test_fallback_update.py TestFallbackUpdate.test_update | IOError: Process kill
07:05FlorinMezeiwhimboo: sure
07:05whimboowhenever you have time, please setup the machine so that I can ssh into
07:05FlorinMezeiI have some tests running for the RC
07:06whimbook, ping me afterward then
07:06FlorinMezeias soon as they are done I'll take the machine offline
07:06FlorinMezeiI also have full logs from yesterday if those help
07:17FlorinMezeiwhimboo: mm-ub-1404-32-3.qa.scl3.mozilla.com is now offline
07:20whimbooFlorinMezei: the log output on the bug was not the pretty helpful
07:20whimbooso lets see
07:20whimboocan you prepare the tests?
07:21FlorinMezeisure
07:22FlorinMezeiready to run
07:25FlorinMezeiI hope :) - whimboo ^
07:26whimboosure you are on mm-ub-1404-32-3?
07:26whimbooi only have one screen window
07:26whimboowhich is the running java client
07:26FlorinMezeihm... yeah pretty sure
07:26FlorinMezeiI see 2 screen windows
07:27whimboowell, you attached a new screen session
07:27whimboos/attached/created
07:27FlorinMezeishould I have used the existing one?
07:27whimboosure
07:27whimboobut anyway
07:27whimbooi will connect to this one now
07:28whimbooFlorinMezei: you are in ondemand_update
07:28whimboowhile it should be running in mozilla-central_update
07:28FlorinMezeiwell that explains a lot
07:29whimboolets test in the other folder
07:29whimbooif it still has issues I will take a look
07:29FlorinMezeiok
07:29FlorinMezeiI'll let you know
07:31whimbooFlorinMezei: oh, you dont have to specify the full path all the time
07:31whimboojust use relative paths
07:31FlorinMezeisorry man, I'm a bit slower with Linux :)
07:32whimboono problem :)
07:32whimbooso ../../mozilla-central_update would have also done it
07:33FlorinMezeiwhimboo: is it a problem that the environment is set from ondemand_update?
07:33whimbooyeah. deactive it
07:33whimbooby running "deactivate"
07:33FlorinMezeiok
07:35FlorinMezeiwhimboo: how do I do that?
07:35whimboodone
07:35FlorinMezeidamn :)
07:35whimboovirtualenv adds a global function
07:35whimboowhich is available from everywhere in the current shell
07:36whimbooit doesn't depend in which path you currently are
07:36whimbooFlorinMezei: btw. as you may have seen I landed the patch to click the restart button for udpate tests
07:36FlorinMezeiindeed
07:36whimbooi'm kinda waiting for the centrla Nightly!
07:39FlorinMezeiwhimboo: starting the test
07:42whimboonow comes the interseting part
07:42whimbooand it works
07:42FlorinMezeiwoohoo
07:42whimbooso, it is only happening for fallback or?
07:43whimbooso specify --update-fallback-only
07:43whimbooit prevents us from running the direct too
07:43FlorinMezeiok
07:43whimbooalso we had a complete fallback now
07:43whimboonot a partial fallback
07:44whimbooso you may have to download a newer nightly as what we currently have
07:44whimboomaybe complete fallback doesn't show this behvior
07:44FlorinMezeiok, let me do that
07:44whimboowhen you have that... lets run this test a dozen of times
07:45whimbooif its all passing, we should also try with a regular central nightly
07:45whimboojust to prove that this fails
07:46FlorinMezeiwhimboo: so I used the second most recent oak: 2017-04-17-11-03-20-oak/
07:46FlorinMezeithere is only on that's more recent: 2017-04-18-11-03-18-oak/
07:46FlorinMezei*one
07:47FlorinMezeioh no wait, maybe it was the 3rm most recent
07:48whimboohm? i thought it downloaded ~65MB twice
07:51FlorinMezeiwhimboo: started again
07:51FlorinMezei2017-04-17-11-03-20-oak/ - fallback only
07:52whimboohm, again a full patch
07:52whimbooi wonder if we dont have partials
07:52whimboolets wait for the final results
07:53whimbooso to scroll up you have to press ctrl+a esc
07:53whimbooto quit the scrolling mode press esc twice
07:54FlorinMezeinice
07:55whimboohm
07:55whimboohttps://aus5.mozilla.org/update/6/Firefox/55.0a1/20170417110320/Linux_x86-gcc3/en-US/nightly-oak/Linux%203.13.0-106-generic%libpulse%204.0.0)/NA/default/default/update.xml?force=1
07:55whimbooisn't that correct?
07:55whimbooI get a bad request
07:55FlorinMezeiI get the same... odd
07:56FlorinMezeithere's a paranthesis missing
07:56FlorinMezeihttps://aus5.mozilla.org/update/6/Firefox/55.0a1/20170417110320/Linux_x86-gcc3/en-US/nightly-oak/Linux%203.13.0-106-generic%20(GTK%203.10.8%2Clibpulse%204.0.0)/NA/default/default/update.xml?force=1
07:57whimbooalso generic%ibpulse
07:57FlorinMezeiindeed
07:57whimboodo I miss a 'l'?
07:58FlorinMezeimy link works fine
07:58FlorinMezeiserves a complete mar
07:58whimboolooks like we only serve complete mar files :/
07:59whimbook, so we would have to test complete mars for mozilla-central then
07:59whimboojust pick a today -5
07:59whimbooso 13th April
07:59whimboogiven that we do not run our update tests for complete, we should check if those are also affected
08:00FlorinMezei2017-04-13-11-02-39-oak/ should do I guess
08:00whimboonot oak
08:00whimboomozilla-central
08:00FlorinMezeioh, centrall
08:00FlorinMezeiok
08:00FlorinMezei2017-04-13-10-01-44-mozilla-central/
08:01whimbooFlorinMezei: i checked latest RC updates
08:01whimboolooks like complete all pass
08:01FlorinMezeiyou mean what I've run today for release-cdntest?
08:02whimbooupdates from 53.0b10 are partials
08:03whimboothose fail
08:03whimboo53.0b9 and earlier are complete
08:03FlorinMezeiah, right
08:03whimbooand do not fail
08:03whimbooso i believe the issue here is really with partial updates
08:04FlorinMezeisounds about right
08:04whimbooand that we cannot verify on oak yet :(
08:04whimbooI will comment on the bug
08:04FlorinMezeiwhimboo: should I still test with the older mozilla central?
08:04whimbooi dont think it makes sense
08:05FlorinMezeiagreed
08:05whimboobring the machine back online and we are done for now
08:05FlorinMezeiok, thanks
08:05whimboolets hope we can get partials. if not no idea
20 Apr 2017
No messages
   
Last message: 5 days and 11 hours ago