743901 - need access to Win7 test slave

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: CIDuty, task, P3)

Description

[John Daggett (:jtd)]

Attempting to land a fix for bug 710727, I'm seeing reftest/mochitest failures on one of the Win7 machines that I can't reproduce locally.

https://tbpl.mozilla.org/php/getParsedLog.php?id=10745705&tree=Mozilla-Inbound&full=1

https://tbpl.mozilla.org/php/getParsedLog.php?id=10745088&tree=Mozilla-Inbound&full=1

I'd like to remote access one of these machines to run my tests there.

Comment 1

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

found in triage.

Comment 2

[John Daggett (:jtd)]

I should add here that the problem only seems to occur one the Win7 *debug* machines.

Comment 3

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to John Daggett (:jtd) from comment #2)
> I should add here that the problem only seems to occur one the Win7 *debug*
> machines.

fyi: We use the same builder machines for building both opt or debug builds. It helps us scale better.

Comment 4

[John Daggett (:jtd)]

This is blocking work (i.e. landing) that I'd like to land by the middle of next week.

It also appears to be a bit of a heisenbug.  It failed yesterday on Win7 debug try but today it succeeds.  Argh.

failures
========

http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-f1bd1195612c/try-win32-debug/try_win7-debug_test-mochitests-4-bm15-tests1-windows-build2821.txt.gz

builder: try_win7-debug_test-mochitests-4
slave: talos-r3-w7-065
starttime: 1334039913.24
results: warnings (1)
buildid: 20120409204039
builduid: a7da5267aec94baa803735bc15147225
revision: f1bd1195612c

http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-a015cc8b7ff4/try-win32-debug/try_win7-debug_test-reftest-bm23-tests1-windows-build655.txt.gz

builder: try_win7-debug_test-reftest
slave: talos-r3-w7-061
starttime: 1334078384.04
results: warnings (1)
buildid: 20120410072900
builduid: 52872476d5c94a15a995218bb55a10b0
revision: a015cc8b7ff4

success (font-matching/font-face):
========

http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-8e26c34f8ca1/try-win32-debug/try_win7-debug_test-reftest-bm15-tests1-windows-build2729.txt.gz

builder: try_win7-debug_test-reftest
slave: talos-r3-w7-006
starttime: 1334113117.65
results: success (0)
buildid: 20120410180403
builduid: f292226e4d494d3e81fc5ab40c2dd53e
revision: 8e26c34f8ca1

http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-5e7f3613bf21/try-win32-debug/try_win7-debug_test-reftest-bm23-tests1-windows-build667.txt.gz

builder: try_win7-debug_test-reftest
slave: talos-r3-w7-048
starttime: 1334131556.6
results: warnings (1)
buildid: 20120410225003
builduid: 75dd85b9376243d4927fa395347f5a16
revision: 5e7f3613bf21

The fallback behavior is slightly bizarre:

REFTEST TEST-START | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/font-face/src-list-local-full.html | 4049 / 7449 (54%)
(fontmatching) font: [Times New Roman Bold] match: systemfallback len: 32 [This should be a sans-serif face]
(fontmatching) font: [Times New Roman Bold] match: systemfallback len: 37 [This should be a bold sans-serif face]
(fontmatching) font: [Times New Roman Bold] match: systemfallback len: 40 [This should be an italic sans-serif face]
REFTEST TEST-START | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/font-face/src-list-local-full-ref.html | 4049 / 7449 (54%)
(fontmatching) font: [Arial Bold] match: systemfallback len: 32 [This should be a sans-serif face]
(fontmatching) font: [Arial Bold] match: fontgroup len: 37 [This should be a bold sans-serif face]
(fontmatching) font: [Arial Italic] match: fontgroup len: 40 [This should be an italic sans-serif face]
REFTEST TEST-UNEXPECTED-FAIL | file:///c:/talos-slave/test/build/reftest/tests/layout/reftests/font-face/src-list-local-full.html | image comparison (==)

Note how no regular face is used!

If at all possible, it would be good to get access to one of the machines that failed (e.g. talos-r3-w7-065 or talos-r3-w7-061).

Comment 5

[Armen [:armenzg]]

Assigning this critical bug to this week's buildduty.

bear, can you have a look at this today?

Thanks.

Comment 6

[Mike Taylor [:bear]]

John,

what is your ldap userid - I am setting up talos-r3-w7-065 as a loaner for you and need to enable your access.

Comment 7

[John Daggett (:jtd)]

ldap: [e-mail]

Comment 8

[Mike Taylor [:bear]]

filed bug for buildvpn access and dropped priority down to major.

host is ready for you once you get access - I need to give you the username and password once you have access.

Comment 9

[Nick Thomas [:nthomas] (UTC+12)]

John was having trouble with the responsiveness of the slave using VNC, so I've tried upgrading UltraVNC to 1.0.9.6.2 (from Feb 2012).

Comment 10

[John Daggett (:jtd)]

The following UltraVNC changes made things bearable (and *possible*!!):

Select Task tray > UltraVNC widget (screen with eyeball) > Right click

Admin properties
Misc
unselect Remove Aero, Remove Wallpaper, Enable Blank Monitor...
select Enable Capture alpha blending
  http://forum.ultravnc.info/viewtopic.php?t=18720

Properties
select Poll Full Screen

The "Enable Capture alpha blending" is important to allow apps that use an Aero look to draw menus correctly.  This includes Explorer, when app windows are grouped clicking the app icon on the task bar will bring up a list of all windows for that app.  Without this option clicked *nothing* will appear, although if you move the mouse and click you can play "pin the tail on the menu".

Comment 11

[John Daggett (:jtd)]

I'm all done, problem solved.  Thanks very much for the use of the slave machine.

Comment 12

[Rail Aliiev [:rail]]

Needs post-reimaging once bug 749614 is done.