[00:00:39] <seanpaul> davidjames: no clue what's happening there... it was churning for a while, presumably on the update? The logs are pretty sparse
[00:00:42] <pstew> I'm not sure whre the failure is.
[00:00:45] *** behdad has quit IRC
[00:01:05] <davidjames> seanpaul: Autoupdate was unsuccessful.  Status returned was UPDATE_STATUS_IDLE.
[00:02:35] <seanpaul> davidjames: hmm... sounds like it might be a legit bug
[00:02:40] <davidjames> seanpaul: Also noticed "Failed to copy syslinux files in and make them bootable"
[00:02:47] <davidjames> seanpaul: Take a look at update.log in test_results.tgz
[00:03:02] <pstew> seanpaul,davidjames: Elsewhere in the PFQ logs I see "FAIL VMTest (0:29:22) in bin/cros_au_test_harness"
[00:03:25] <davidjames> pstew: Yup, that tells you the test fails, but doesn't tell you why -- update.log says why
[00:03:29] <seanpaul> davidjames: that log line rings a bell (failed to copy syslinux files)
[00:06:13] <davidjames> seanpaul: Looks like chromeos-setimage failed. Ellyjones just updated chromeos-setimage
[00:06:41] <davidjames> seanpaul: http://google.com/codesearch#wZuuyuB8jKQ/src/platform/installer/chromeos-chroot-postinst&exact_package=chromiumos&q=%22Failed%20to%20copy%20syslinux%22&l=263
[00:07:08] <seanpaul> davidjames: I was just looking at this
[00:07:28] <seanpaul> ellyjones: you have a change in the installer repo 90 minutes ago, any of this look familiar?
[00:07:36] <davidjames> seanpaul: Yeah you can see from above search that that error message is printed any time chromeos-setimage fails
[00:07:57] <ellyjones> hmm
[00:08:04] <ellyjones> it is possible, let me check
[00:08:12] <seanpaul> ellyjones: thanks
[00:08:56] <crosbot> tree became 'Tree is closed (crosbug.com/20034 strikes again->nirnimesh, x86-pfq failing on update script->ellyjones investigating)'
[00:09:08] <davidjames> seanpaul: Since x86-alex-binary is passing that suggests that maybe a cros-workon ebuild is causing the failure
[00:09:09] <ellyjones> (my change is behind a flag that is not set)
[00:09:34] <davidjames> seanpaul: (If it's not a flaky failure, it's likely something that was just committed that the PFQ is revving)
[00:09:52] <seanpaul> ellyjones: alright, i'll keep digging, thanks
[00:11:26] <seanpaul> davidjames: maybe it's flakey, I remember kliegs saying something about that log line this morning, but I was only half listening (damn!)
[00:12:07] <davidjames> seanpaul: verity '' segfaults for me
[00:12:09] <ellyjones> ah, no, it is my fault -.-
[00:12:12] * ellyjones revert
[00:12:26] <ellyjones> or I could fix verity
[00:12:41] <seanpaul> ellyjones: good news! how long to fix?
[00:12:59] <ellyjones> probably a few minutes
[00:13:21] <seanpaul> I'm fine with that, but my sheriffing shift is over :)
[00:13:29] <pstew> Do we want to revert in the mean time so you don't feel rushed in testing?
[00:13:51] <ellyjones> the fix is very small, I believe
[00:14:05] <pstew> Okay.
[00:14:29] <seanpaul> ellyjones, pstew: We might as well fix it since nirnimesh is still bandaging the test, yeah?
[00:15:02] <ellyjones> fix is up at http://gerrit.chromium.org/gerrit/7294
[00:15:35] <ellyjones> that should've been there all along since the rest of the arg parsing assumes key and val are set
[00:16:30] <ellyjones> anyone want to review? :)
[00:16:44] <pstew> Looking.
[00:16:52] <seanpaul> ellyjones: LGTM, but I'll let someone else +2 it since I've never touched that part of the codebase
[00:17:10] <ellyjones> cool
[00:17:13] <davidjames> ellyjones: It's also a bug to pass in null arguments, maybe not pass in null arguments? :)
[00:17:22] <ellyjones> davidjames: it makes verity easier to program around
[00:17:41] <ellyjones> see http://gerrit.chromium.org/gerrit/#change,7275 - if verity doesn't take null arguments, I have to have two separate invocations of it
[00:18:09] <seanpaul> pstew, robotboy_google: I'm off for the day, pleasure working with you guys :)
[00:18:11] <pstew> ellyjones: Um, TEST=?
[00:18:26] <ellyjones> pstew: Adhoc, yeah?
[00:18:42] <ellyjones> the line below is the test :P
[00:18:54] <gauravsh> I tried an Adhoc test with your change and it failed. :)
[00:19:02] <gauravsh> since my test was adhoc, i don't remember what it was.
[00:19:13] <pstew> Oh dear.  I need to re-learn scrollbars? :-)
[00:20:14] <ellyjones> the line below says what my test was, though :P
[00:21:18] <davidjames> ellyjones: What happens if an empty value is passed in intentionally? (E.g. to set some key to the empty string)
[00:21:44] <ellyjones> davidjames: that isn't allowed (right now) and verity would've segfaulted on it before
[00:21:48] <ellyjones> now it just skips that key
[00:22:24] <ellyjones> anyway, the VM tests should now not be combusted
[00:23:08] <gauravsh> ellyjones: it might be useful to log a warning when a key is skipped.
[00:23:54] <davidjames> ellyjones: Segfault might be better than success, since it indicates the user maybe did something wrong :)
[00:24:06] <ellyjones> gauravsh: yeah
[00:24:08] <ellyjones> davidjames: :)
[00:24:23] <ellyjones> empty value may be user error, but empty key is intentionally supported
[00:24:56] <davidjames> ellyjones: Yeah, I agree, empty key is OK but empty value might actually be a real bug in the calling code
[00:25:27] <nirnimesh> robotboy: sent you CL for bandaging 20034
[00:25:28] <ellyjones> davidjames: I filed a bug to remind me to do that
[00:25:33] <davidjames> Agreed that an error message would be more friendly than a segfault :)
[00:25:49] <davidjames> ellyjones: thanks
[00:27:35] <davidjames> nirnimesh: Is NavigateToUrl blocking? I.e. does it just ask browser to navigate the specified page, or does it wait until specified page is fully loaded?
[00:27:46] <nirnimesh> davidjames: Yes, it's blocking
[00:28:18] *** saintlou has quit IRC
[00:28:21] <nirnimesh> over 300 other pyauto tests depend on that, and they're fine
[00:28:52] <davidjames> nirnimesh: Hmm but are any of the 300 other tests running in smoke suite?
[00:29:16] <nirnimesh> davidjames: no
[00:30:05] <davidjames> nirnimesh: Also, another question, how do you know the httpserver has fully started by the time chrome goes to connect to it?
[00:31:09] <nirnimesh> davidjames: httpd.py takes care of that. A lot of tests already use httpd
[00:31:44] <nirnimesh> davidjames: but if it's really that, my bandaging CL will be able to tell
[00:32:03] <nirnimesh> davidjames: available for review at http://gerrit.chromium.org/gerrit/#change,7295
[00:32:18] <davidjames> nirnimesh: HTTPListener does not guarantee server is started
[00:32:20] <oshima> nirnimesh: how/when pyauto_dep file is created? --use_emerged dodm
[00:32:38] <nirnimesh> oshima: it's created when you emerge chromeos-chrome
[00:32:45] <oshima> stopped working after I created new vm image
[00:33:04] <davidjames> nirnimesh: We also had bugs in crostestutils with similar code, the race condition was hitting where we were 'really fast' at connecting to the specified server before it had started
[00:33:26] <nirnimesh> davidjames: I see.
[00:33:29] <oshima> nirnimesh: ok, so I need to emerge chromeos-chrome after I build the image?
[00:33:40] <davidjames> nirnimesh: Take a look at the code there, it just launches a thread
[00:33:57] <nirnimesh> oshima: do you build chrome from source?
[00:33:59] <davidjames> nirnimesh: In other versions of the same class we added a semaphore to make sure server was fully started
[00:34:04] <davidjames> nirnimesh: No such code there
[00:34:14] <oshima> nirnimesh: yes
[00:34:17] <nirnimesh> davidjames: hmm.. example?
[00:35:39] <davidjames> nirnimesh: http://gerrit.chromium.org/gerrit/4778
[00:35:47] <davidjames> nirnimesh: Not the exact same class, but similar issues
[00:36:10] <robotboy_google> nirnimesh: I don't know pyauto, but it looks like that would generate a lot of log output until the title matches.  What is the poll frequency of poll_for_condition?
[00:36:19] *** TW1920_ has quit IRC
[00:38:44] <nirnimesh> robotboy_google: poll_for_condition is not a pyauto thing. sleep_interval=0.1 secs
[00:39:24] <nirnimesh> robotboy_google: this code is temporary. What davidjames says seems correct, so the http server needs to be fixed to be reliable.
[00:40:07] <davidjames> nirnimesh: Yes I guess problem here is that nobody expected Chrome to be so fast that it would outpace the HTTPServer class ;)
[00:40:17] *** saintlou has joined #chromium-os
[00:40:17] *** ChanServ sets mode: +v saintlou
[00:40:51] <nirnimesh> davidjames: yeah, which is why this problem keeps getting clubbed with pyauto. pyauto allows talking to chrome really really quickly after launch
[00:41:00] <ellyjones> Break Tests Chrome Fast.
[00:42:03] <robotboy_google> OK, looks good to me.  But I feel like someone else should give it a +2 considering I've never worked in the code before.
[00:42:06] <ellyjones> are you guys talking about http://code.google.com/p/chromium-os/issues/detail?id=19828 ?
[00:44:12] <nirnimesh> ellyjones: yes. same thing. had disappeared for a while (maybe chrome got slow :P)
[00:46:45] <nirnimesh> davidjames: can you look at http://gerrit.chromium.org/gerrit/#change,7094 ? I'll look into fixing httpd.py later but for now the bandaging should keep the tree open
[00:47:27] <davidjames> nirnimesh: Yeah almost looks good, can you add in some TODOs?
[00:47:32] <davidjames> nirnimesh: And update the bug link to be correct? :)
[00:47:39] <davidjames> nirnimesh: review feedback is inline
[00:49:04] <robotboy_google> davidjames: Thanks for the additional review.
[00:57:01] *** saintlou has quit IRC
[00:59:35] *** saintlou has joined #chromium-os
[00:59:35] *** ChanServ sets mode: +v saintlou
[01:00:30] *** magn3ts has joined #chromium-os
[01:01:30] *** sbyer has joined #chromium-os
[01:01:31] *** ChanServ sets mode: +v sbyer
[01:10:46] <nirnimesh> sheriffs: permission to land http://gerrit.chromium.org/gerrit/7295 to stop-gap fix crosbug.com/20034 please
[01:11:18] <robotboy_google> Yup, go for it.  Thanks!
[01:12:44] *** wfrichar has quit IRC
[01:13:08] <nirnimesh> Thanks
[01:15:10] <crosbot> tree became 'Tree is closed (dektopui_UrlFetch fix landed, x86-pfq failing on update script->ellyjones investigating)'
[01:17:16] *** vmil86 has quit IRC
[01:17:41] *** jochen___ has joined #chromium-os
[01:17:42] *** ChanServ sets mode: +v jochen___
[01:21:02] *** jochen__ has quit IRC
[01:21:03] *** jochen___ is now known as jochen__
[01:21:37] *** sadrul has quit IRC
[01:28:58] <pstew> Okay.  I'm un-reporting for duty.  My shuttle is soon and connectivity (and elbow room) can be limited from there.
[01:29:42] <eggy> heh
[01:32:26] <robotboy_google> pstew: See ya
[01:34:26] *** behdad has joined #chromium-os
[01:40:41] *** lipsinV2 has joined #chromium-os
[01:43:49] *** behdad has quit IRC
[01:50:14] *** patcito has quit IRC
[01:53:02] *** sbyer has quit IRC
[01:54:17] <robotboy_google> signing off, heading for my shuttle.
[01:55:25] *** robotboy_google has left #chromium-os
[01:55:32] *** saintlou is now known as saintlou|away
[01:59:02] *** stevenjb has quit IRC
[01:59:26] <oshima> cmasone: ping
[01:59:34] *** pstew_mobile_ has joined #chromium-os
[02:04:07] *** sbyer has joined #chromium-os
[02:04:07] *** ChanServ sets mode: +v sbyer
[02:09:33] *** vbendeb has joined #chromium-os
[02:09:34] *** ChanServ sets mode: +v vbendeb
[02:09:44] <davidjames> PFQ looks good, suggesting reopening
[02:10:24] <nirnimesh> I wish I could see logs as tests run
[02:10:36] <nirnimesh> oshima: cmasone is on vacation
[02:13:15] *** saintlou|away is now known as saintlou
[02:15:20] *** chocobo__ has quit IRC
[02:16:07] <vbendeb> I wonder if any of the sheriffs are still around to open the tree...
[02:18:41] <davidjames> vbendeb: Nope, all sheriffs went home
[02:18:49] <gauravsh> I think they are all out. Channel is sheriff.
[02:19:06] <gauravsh> ellyjones: are you still looking at the x86-pfq failure?
[02:19:09] <davidjames> vbendeb: Oh, note though that PFQ is failing due to different issue
[02:19:28] <davidjames> vbendeb: android.git.kernel.org is down, hence repo sync is failing :(
[02:20:13] *** saintlou has quit IRC
[02:20:16] <gauravsh> wasn't there a PSA to use the gerrit mirror instead of kernel.org?
[02:20:59] <davidjames> gauravsh: Yeah, rcui updated the buildbots to use the gerrit mirror, but the update didn't work
[02:21:33] <crosbot> tree became 'Tree is closed (updating buildbots to use gerrit mirror for repo -> rcui)'
[02:21:46] <davidjames> On the plus side, tests are passing!
[02:22:27] <vbendeb> guys, I have a u-boot patch withc is not going to affect any of the builders (it is used on x86 onlty, and builders don't build x86 u-boot)
[02:22:41] <vbendeb> other people need this to proceed...
[02:22:52] <davidjames> vbendeb: Sure that's fine to go in
[02:23:02] <vbendeb> davidjames: ok
[02:23:15] <davidjames> vbendeb: Tree is basically green except bots can't sync properly, no actual problems with the build
[02:23:28] *** dennisjeffrey has quit IRC
[02:24:06] *** chocobo__ has joined #chromium-os
[02:24:06] *** ChanServ sets mode: +v chocobo__
[02:28:43] *** chocobo__ has quit IRC
[02:29:00] <nirnimesh> tests passing. Yay! So my webui login automation did work. xiyuan/zel must be pleased
[02:33:58] *** stevenjb has joined #chromium-os
[02:34:32] *** rbyers_ has joined #chromium-os
[02:39:10] *** rbyers_ has quit IRC
[02:39:17] *** rbyers_ has joined #chromium-os
[02:39:17] *** ChanServ sets mode: +v rbyers_
[02:45:57] *** deshantm has joined #chromium-os
[02:47:56] *** saintlou has joined #chromium-os
[02:47:56] *** ChanServ sets mode: +v saintlou
[02:49:13] *** saintlou has quit IRC
[02:56:40] *** patcito has joined #chromium-os
[03:14:20] <redpig> win 21
[03:14:23] *** kochi has quit IRC
[03:14:24] <redpig> doh
[03:19:35] *** Adys has quit IRC
[03:22:13] *** pstew_mobile_ has quit IRC
[03:23:00] *** kochi has joined #chromium-os
[03:23:10] *** kochi has quit IRC
[03:23:26] *** kochi has joined #chromium-os
[03:39:29] *** rbyers_ has quit IRC
[03:43:57] *** sadrul has joined #chromium-os
[03:43:58] *** ChanServ sets mode: +v sadrul
[03:57:31] *** deshantm has quit IRC
[03:57:32] *** msb__ has quit IRC
[03:57:32] *** davidjames has quit IRC
[03:57:32] *** Calchan has quit IRC
[03:59:32] *** deshantm has joined #chromium-os
[03:59:32] *** msb__ has joined #chromium-os
[03:59:32] *** davidjames has joined #chromium-os
[03:59:32] *** Calchan has joined #chromium-os
[03:59:32] *** pratchett.freenode.net sets mode: +vv msb__ davidjames
[04:00:41] <crosbot> tree became 'Tree is open (rcui->most builders have repo url updated.  Clobber if android.kernel.org error appears.)'
[04:19:49] *** BThompson has joined #chromium-os
[04:19:50] *** ChanServ sets mode: +v BThompson
[04:22:00] *** BThompson has left #chromium-os
[04:22:42] *** BThompson has joined #chromium-os
[04:22:42] *** ChanServ sets mode: +v BThompson
[04:22:47] *** BThompson has quit IRC
[04:22:55] *** BThompson has joined #chromium-os
[04:22:55] *** ChanServ sets mode: +v BThompson
[04:23:18] *** BThompson has quit IRC
[04:23:25] *** BThompson has joined #chromium-os
[04:23:25] *** ChanServ sets mode: +v BThompson
[04:23:42] *** BThompson has quit IRC
[04:23:51] *** BThompson has joined #chromium-os
[04:23:51] *** ChanServ sets mode: +v BThompson
[04:44:20] *** corburn has joined #chromium-os
[05:09:41] *** stevenjb has quit IRC
[05:10:26] *** stevenjb has joined #chromium-os
[05:38:04] *** Inumedia has quit IRC
[05:38:33] *** Inumedia has joined #chromium-os
[05:41:20] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from 2047f9ade3b1c9e13494e1a0f8927a9f4a877e23: _third_party_ at chromium dot org)'
[05:44:43] *** pstew_mobile_ has joined #chromium-os
[05:45:34] <pstew_mobile_> Howdie folks.  Just checking in.  Can't really take on sheriffing duties from here, but was wondering whether folks were getting along with the chrome build-fail i the PFQ.
[05:56:27] <crosbot> tree became 'Tree is closed (yjlou is looking. TOT PFQ login_BadAuthentication failed. )'
[06:08:39] *** pstew_mobile_ has quit IRC
[06:15:55] <crosbot> tree became 'Tree is open (filed TreeCloser chromium-os:20109)'
[06:16:56] <crosbot> tree became 'Tree is open (filed issue chromium-os:20109)'
[06:45:19] *** chocobo__ has joined #chromium-os
[06:45:20] *** ChanServ sets mode: +v chocobo__
[06:49:13] *** chocobo___ has joined #chromium-os
[06:49:13] *** ChanServ sets mode: +v chocobo___
[06:51:32] *** chocobo__ has quit IRC
[06:51:33] *** chocobo___ is now known as chocobo__
[07:07:15] *** gfrog has quit IRC
[07:08:23] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "x86-alex_he canary" from None: )'
[07:09:10] *** gfrog has joined #chromium-os
[07:10:49] *** gfrog has quit IRC
[07:15:02] *** gfrog has joined #chromium-os
[07:24:33] *** reveman has quit IRC
[07:24:33] *** dots has quit IRC
[07:24:33] *** Kyngdom has quit IRC
[07:24:33] *** pastarmovj has quit IRC
[07:24:33] *** chubb_cloud has quit IRC
[07:24:33] *** Hexxeh has quit IRC
[07:39:08] *** stevenjb has quit IRC
[07:42:56] *** petermayo has joined #chromium-os
[07:42:57] *** ChanServ sets mode: +v petermayo
[07:43:49] <petermayo> does anyone here know if the _rc and _alpha suffixes we put on ebuilds have semantic meaning to portage?
[07:56:38] *** Hexxeh has joined #chromium-os
[07:58:26] *** sergiu has quit IRC
[08:07:39] *** kinaba has quit IRC
[08:14:02] *** nona has quit IRC
[08:21:39] <crosbot> tree became 'Tree is open (looks like random crashes in suite_Smoke test cases)'
[08:26:57] *** kinaba has joined #chromium-os
[08:28:55] *** Styx has joined #chromium-os
[08:31:35] *** kinaba has quit IRC
[08:32:52] *** mnissler has quit IRC
[08:33:07] *** mnissler has joined #chromium-os
[08:33:07] *** ChanServ sets mode: +v mnissler
[08:38:07] *** kinaba has joined #chromium-os
[08:41:56] *** Hexxeh has quit IRC
[09:07:34] *** reveman has joined #chromium-os
[09:08:54] *** dots has joined #chromium-os
[09:10:30] *** Kyngdom has joined #chromium-os
[09:12:24] *** chubb_cloud has joined #chromium-os
[09:15:16] *** petermayo has left #chromium-os
[09:16:24] *** pastarmovj has joined #chromium-os
[09:20:06] *** Hexxeh has joined #chromium-os
[09:26:10] *** gfrog has quit IRC
[09:32:47] *** gfrog has joined #chromium-os
[09:33:51] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from f8a83e23f658b918edb17a77b0e0ac8a9335d73f: dbeam at chromium dot org <dbeam at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[09:34:13] *** mnissler_ has joined #chromium-os
[09:34:13] *** ChanServ sets mode: +v mnissler_
[09:36:28] *** corburn has quit IRC
[09:38:03] *** mnissler has quit IRC
[09:40:44] *** chocobo__ has quit IRC
[09:50:29] <crosbot> tree became 'Tree is open (filed issue chromium-os:20109)'
[09:58:02] *** chocobo__ has joined #chromium-os
[09:58:03] *** ChanServ sets mode: +v chocobo__
[09:58:23] *** gfrog has quit IRC
[09:58:24] *** chocobo___ has joined #chromium-os
[09:58:25] *** ChanServ sets mode: +v chocobo___
[10:02:24] *** gfrog has joined #chromium-os
[10:03:13] *** chocobo__ has quit IRC
[10:05:10] *** chocobo___ has quit IRC
[10:11:12] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "stumpy canary" from None: )'
[10:11:51] *** patcito has quit IRC
[10:28:01] *** JakeSays has quit IRC
[10:29:18] *** JakeSays has joined #chromium-os
[10:31:47] *** robe_ has quit IRC
[10:38:20] *** robe_ has joined #chromium-os
[10:44:29] <crosbot> tree became 'Tree is open (Resolving host timed out: www.uk.kernel.org, not CL broke the tree)'
[11:03:59] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot" on "tegra2_seaboard-tangent canary" from None: )'
[11:24:28] *** gfrog has quit IRC
[11:36:23] *** gfrog has joined #chromium-os
[11:41:36] <crosbot> tree became 'Tree is throttled (tree keeps broken due to non-CL issues: kernel.org timeout and random crashes in smoke tests)'
[11:53:24] *** TW1920 has joined #chromium-os
[11:53:35] *** achuith has joined #chromium-os
[11:53:35] *** ChanServ sets mode: +v achuith
[11:57:08] *** vmil86 has joined #chromium-os
[12:33:43] *** Styx has quit IRC
[12:35:49] *** Styx has joined #chromium-os
[13:07:07] *** borios has joined #chromium-os
[13:22:46] *** borios_ has joined #chromium-os
[13:22:47] *** borios has quit IRC
[13:33:00] *** gfrog has quit IRC
[13:39:19] *** gfrog has joined #chromium-os
[13:51:40] *** borios has joined #chromium-os
[13:51:41] *** borios_ has quit IRC
[14:12:56] *** vpalatin|home has joined #chromium-os
[14:14:31] <vpalatin|home> As I'm still on the east coast shift, I'm sheriffing a little bit from home before going to the office
[14:21:58] *** sadrul has quit IRC
[14:39:37] *** hru has joined #chromium-os
[14:41:57] *** yoshiki_ has quit IRC
[14:42:23] *** yoshiki_ has joined #chromium-os
[14:45:03] <hru> i got below error while launching chrome(my own chromium)
[14:45:06] <hru> chrome: /usr/lib/gcc/i686-pc-linux-gnu/4.4.3/gcc/i686-pc-linux-gnu/4.4.3/libstdc++.so.6: version `GLIBCXX_3.4.14' not found (required by ./chrome)
[14:45:06] <hru> chrome: /usr/lib/libjpeg.so.62: version `LIBJPEG_6.2' not found (required by ./chrome)
[14:45:22] <hru> on chrome-os
[14:45:48] <hru> i following procedure mentioned in https://sites.google.com/a/chromium.org/dev/developers/how-tos/build-instructions-chromeos
[14:47:03] *** gfrog has quit IRC
[14:50:46] <vpalatin|home> hru: did you build your chromium inside the chroot ?
[14:51:37] *** SoulRaven has joined #chromium-os
[14:52:29] <hru>  vpalatin|home:no
[14:53:33] <vpalatin|home> hru: it is the easiest way to get it right. here your host glibc version doesn't match the one used by chromeOS
[14:55:57] <hru> can I follow the step that mentioned in above link to build the chrome in chroot ? Or is it the other way ?
[14:58:46] <vpalatin|home> Yes you can use the above instructions if you have a working chromiumos chroot
[14:59:28] <hru> Thanks,
[15:03:08] <vpalatin|home>  I'm heading to the office to dig further the current build mess, I'm off for the next 20min
[15:03:18] *** vpalatin|home has quit IRC
[15:22:24] <vpalatin> back online, I'm investigating what I can do for the udev package
[15:44:56] <vpalatin> benchan, adlr: I have uploaded a copy of udev-170.tar.bz2 to chromeos-localmirror (as kernel.org is down), I'm trying a new build on x86-generic-full
[15:45:22] *** rbyers_ has joined #chromium-os
[15:45:22] *** ChanServ sets mode: +v rbyers_
[15:49:49] *** snickersnack has joined #chromium-os
[15:50:41] *** snickersnack has quit IRC
[16:07:17] *** borios has quit IRC
[16:07:23] *** stevenjb has joined #chromium-os
[16:08:09] *** stevenjb has quit IRC
[16:12:47] *** seumas has joined #chromium-os
[16:12:47] *** ChanServ sets mode: +v seumas
[16:21:56] <ellyjones> cool
[16:21:58] <ellyjones> thanks, vpalatin :)
[16:22:40] *** ers has joined #chromium-os
[16:22:40] *** ChanServ sets mode: +v ers
[16:26:25] <vpalatin> aaarggh this x86-generic-full now fails on chrome build
[16:26:52] <vpalatin> everything is still broken (thus differently) I will close the tree
[16:30:48] <crosbot> tree became 'Tree is closed (workaround for kernel.org, but crashes in smoke tests and chrome build failure in x86-generic-full)'
[16:31:35] *** FX80 has quit IRC
[16:34:44] <ellyjones> alright
[16:34:53] <ellyjones> I'll look at the smoke failures
[16:35:05] *** FX80 has joined #chromium-os
[16:35:05] <ellyjones> woah, _all_ the bots are red
[16:36:05] <vpalatin> ellyjones: the redness was due to the missing udev archive from kernel.org
[16:36:19] <vpalatin> I hope part of it is solved now
[16:36:45] <ellyjones> cool
[16:36:48] <vpalatin> but the x86-generic-PFQ is failing on a the smoke test
[16:37:00] <vpalatin> might be the yesterday login change
[16:37:18] <ellyjones> well, it's failing on login_BadAuthentication
[16:37:42] <ellyjones> so that seems like a reasonable supposition
[16:38:13] <vpalatin> It's assigned to nirnimesh ( http://code.google.com/p/chromium-os/issues/detail?id=20109 )
[16:42:51] <crosbot> tree became 'Tree is open (workaround for kernel.org DNS issue, still frequent crash in smoke tests crosbug.com/20109)'
[16:57:04] *** seventh has joined #chromium-os
[17:03:18] <ellyjones> the thing where cros/master in kernel is not the real master is aggravating
[17:04:09] *** rbyers_ has quit IRC
[17:04:37] <ellyjones> I keep accidentally rebasing to cros/master and then it takes like 15 seconds for me to be like 'hey, wait...'
[17:09:33] *** sadrul has joined #chromium-os
[17:09:34] *** ChanServ sets mode: +v sadrul
[17:12:13] <kliegs> ellyjones: I just got in the habit of always using m/master
[17:12:23] <kliegs> its got an extra bonus of being less leters
[17:13:57] *** petermayo has joined #chromium-os
[17:13:58] *** ChanServ sets mode: +v petermayo
[17:19:55] <vpalatin> some of the trees are turning green, alleluia!
[17:19:59] *** stevenjb has joined #chromium-os
[17:20:16] <ellyjones> woo
[17:22:34] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from 99ff9934eb0e343160b79f63d314f70012ac877f: joi at chromium dot org <joi at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[17:22:50] *** wfrichar has joined #chromium-os
[17:22:51] *** ChanServ sets mode: +v wfrichar
[17:24:06] <crosbot> tree became 'Tree is open (crosbug.com/20109 beats us again)'
[17:29:15] * benchan starts his sheriff duty with a reasonably green tree
[17:31:59] <vpalatin> benchan: when adlr will be here, we need to decide what to do with 20109. It would nice if we can grab nirnimesh  as he is the assignee of the bug
[17:32:21] <benchan> vpalatin: SGTM
[17:38:32] *** Adys has joined #chromium-os
[17:41:49] <adlr> xiyuan: can you look into http://code.google.com/p/chromium-os/issues/detail?id=20109 ?
[17:51:26] <adlr> benchan: I'm in effectively one long meeting from 11:30 to 1pm today, so hopefully you can watch the tree during that
[17:51:49] <benchan> adlr: sure, no problem
[17:51:56] <adlr> benchan: thanks
[17:52:05] <adlr> it's a killer w/ lunch time :(
[17:53:11] <vpalatin> adlr: as a jetlagged east coast sheriff, I can take care
[17:53:27] <adlr> vpalatin: you are sheriff??
[17:53:39] <redpig> crosbot: sheriffs?
[17:53:39] <crosbot> redpig: sheriffs: benchan, adlr, vpalatin, yjlou
[17:53:53] <vpalatin> adlr: yes ... on non west coast shift
[17:54:03] <vpalatin> crosbot is faster than me
[17:54:05] <adlr> vpalatin: okay
[17:54:57] <vpalatin> it takes some time to change from non west to west coast, I asked end of june, I should hopefully be in the right shift for the next round
[17:58:17] *** saintlou has joined #chromium-os
[17:58:17] *** ChanServ sets mode: +v saintlou
[18:12:33] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "x86 generic PFQ" from 4ecd9c66cc4f0d63c1fc5fd54e8a06d97ae21610: Denis Glotov <glotov at chromium dot org>)'
[18:13:35] <crosbot> tree became 'Tree is open (crosbug.com/20109 beats us again)'
[18:13:36] *** lipsinV2 has quit IRC
[18:14:20] <adlr> vpalatin: how did you know it was 20109 again?
[18:14:31] *** petermayo has quit IRC
[18:14:46] <adlr> oh, the vm test log, i see
[18:15:41] <adlr> does anyone know why 'cbuildbot_master' was red, but it kept going?
[18:15:50] <benchan> adlr, vpalatin: seems like sig 11 in CryptohomeMounted this time
[18:16:27] <adlr> ah, yes
[18:16:57] <vpalatin> oops, I missed that
[18:17:59] <adlr> how can we see which version of chrome was running?
[18:18:33] <benchan> adlr: i think it's chromeos-chrome-15.0.871.0_rc-r1 from the build log
[18:19:05] <adlr> ah, so then, was that the version we pinned?
[18:19:06] <seanpaul> benchan,adlr,vpalatin: 15.0.871 is the version we pinned yesterday since 15.0.872 was misbehaving
[18:19:14] <adlr> i see
[18:20:20] <seanpaul> it was passing all of the vmtests pretty reliably, so I think you can probably rule it out as a bad build
[18:21:28] <adlr> can we get access to the log file?
[18:21:41] <adlr> "Generated stack trace for dump /tmp/cbuildbot0I27hY/test_harness/SimpleTestUpdateAndVerify/2_verify/suite_Smoke/login_CryptohomeMounted/sysinfo/var/spool/crash/chrome.20110907.090353.11454.dmp"
[18:22:03] <adlr> s/log/backtrace/
[18:22:57] <seanpaul> adlr: https://sandbox.google.com/storage/?arg=chromeos-image-archive/x86-generic-pre-flight-queue/0.15.1004.0-a1-b731#chromeos-image-archive%2Fx86-generic-pre-flight-queue%2F0.15.1004.0-a1-b731 ?
[18:23:27] * adlr takes a look
[18:23:49] <vpalatin> adlr: you can get it by using the [Artifacts] link in the waterfall
[18:35:15] <benchan> vpalatin, adlr: looks like VMTest caught sig 11 in various tests, probably due to the some crash in chrome?
[18:39:38] <adlr> hm
[18:43:10] *** saintlou is now known as saintlou|away
[18:50:39] <adlr> benchan: did you see that on other builders, or earlier in x86 generic PFQ?
[18:52:38] <benchan> adlr: looking into the logs
[18:56:22] <adlr> so i guess LKGMCandidateSyncCompletion failure doesn't close the tree?
[18:56:36] <benchan> adlr: haven't seen a similar crash on arm/terga, could it be x86 specific?
[18:57:45] <adlr> benchan: perhaps. i didn't look toooo hard, but i've only seen the login_CryptohomeMounted fail the one time on x86 generic pfq
[19:01:54] <benchan> aldr: x86 pineview full had a sig 11 in suite_Smoke/login_BadAuthentication, not sure if it's due to the same crash.  x86 generic chrome / TOT chrome PFQ show another kind of failure
[19:02:28] <adlr> yeah, that's what i see (various types of failure)
[19:03:32] <adlr> when chrome crashes, do we get a backtrace? If so, I haven't figured out how to get it from the artifacts
[19:04:08] <benchan> adlr: I see various core dump files in test_results.tgz
[19:04:55] <benchan> adlr: but I don't know where to get the debug sym for chrome
[19:04:57] <adlr> what path in there?
[19:05:29] <benchan> adlr: https://sandbox.google.com/storage/chromeos-image-archive/x86-generic-pre-flight-queue/0.15.1004.0-a1-b731/test_results.tgz
[19:05:40] <adlr> i mean, path within the tarball
[19:06:31] <benchan> adlr: ./test_harness/SimpleTestUpdateAndVerify/2_verify/suite_Smoke/login_CryptohomeMounted/sysinfo/var/spool/crash/chrome.20110907.090353.11454.dmp
[19:07:05] <adlr> ah ha
[19:08:06] <benchan> adlr: looks like some bad address: #0  0x000000ac in ?? ()
[19:08:46] <adlr> probably a null vtable pointer?
[19:09:08] <adlr> symbols would be awesome
[19:09:14] <benchan> does anyone know if buildbot archives the debug version of chrome?
[19:13:48] <davidjames> vpalatin: So x86 generic PFQ is looking pretty red still
[19:14:15] <davidjames> benchan: PFQ doesn't archive debug symbols, but other bots do
[19:14:27] <adlr> dang. we should fix that
[19:14:47] <davidjames> adlr: So x86 generic full has debug tgz with it
[19:14:58] <davidjames> adlr: Symbols can be generated with that crash
[19:15:03] <adlr> ah
[19:15:19] <davidjames> adlr: https://sandbox.google.com/storage/chromeos-image-archive/x86-generic-full/0.15.1003.0-a1-b314
[19:16:43] <davidjames> adlr: I think the tree should be closed since all bots that run tests are red
[19:16:49] * adlr downloads debug.tgz. in the mean time, is there a reference on how to get the symbolized crash dump going?
[19:17:11] <adlr> like, do i need to run some command, or is it already in a nice TXT fil?
[19:17:12] <adlr> file
[19:17:34] <davidjames> adlr: http://code.google.com/p/chromium-os/issues/detail?id=17898#c29
[19:18:02] <adlr> thanks
[19:18:59] <vpalatin> adlr, benchan: so we close or we don't ?
[19:19:33] <adlr> i'm fine with a close. to be honest i don't know what the issue really is yet
[19:20:36] <benchan> adlr, vpalatin: i'm fine with a close too
[19:20:50] <vpalatin> Ok I will close
[19:21:20] <crosbot> tree became 'Tree is closed (need to sort out crosbug.com/20109 and other VMtest crashes)'
[19:21:52] *** behdad has joined #chromium-os
[19:28:09] <benchan> cp chrome.debug to VM: No space left on device :(
[19:28:27] <davidjames> benchan: Yeah no surprise, chrome.debug is huge
[19:28:43] <davidjames> benchan: You can either make a bigger image, or don't copy the debug symbols to the VM
[19:29:13] <davidjames> benchan: BTW why are you copying debug symbols to the VM? you can generate stack traces just fine outside the VM
[19:29:52] <benchan> davidjames: oh yes, it's x86 binary
[19:30:27] <davidjames> benchan: Yeah also minidump_stackwalk lives in the chroot so you can run it outside, I think adlr is working on that now
[19:31:01] <davidjames> benchan: If you look at buildbot logs you'll also spot that it tried to run minidump stackwalk but bumped into a sudo bug  which thieule is working on fixing
[19:32:37] <adlr> find test_harness -name "chrome*.dmp" ---> nothing
[19:32:56] <davidjames> adlr: Are there any *.dmp files in there?
[19:33:05] <davidjames> adlr: If there's no dmp files that means no crash
[19:33:11] <adlr> just crasher_nobreakpad
[19:33:18] <davidjames> adlr: Oops I guess that one didn't crash
[19:33:22] <benchan> aldr: I see a few dmp files
[19:33:58] <davidjames> benchan: Yeah adlr was looking at x86 generic full which in fact didn't crash, sorry, just failed
[19:34:00] *** patcito has joined #chromium-os
[19:34:05] *** rbyers_ has joined #chromium-os
[19:34:21] <davidjames> Unhandled JSONInterfaceError: Automation call {'username': 'performance.test.account at googleapps dot com', 'password': 'perfsmurf', 'command': 'Login'} received empty response.  Perhaps the browser crashed.
[19:34:38] <benchan> davidjames: is *.dmp.txt the minidump_stackwalk output?
[19:34:42] <adlr> benchan: I was lookinag at the last x86-generic-full failure
[19:34:45] <adlr> benchan: https://sandbox.google.com/storage/?arg=chromeos-image-archive/x86-generic-full/0.15.1003.0-a1-b314#chromeos-image-archive%2Fx86-generic-full%2F0.15.1003.0-a1-b314
[19:35:04] <davidjames> benchan: Yes but it is not symbolized there, if you run it yourself you'll probably get symbolized output (assuming you have debug symbols)
[19:35:43] <davidjames> benchan: Yeah so for x86-preflight you could grab debug symbols from some build that includes same bit-for-bit identical version of Chrome (NOT x86 generic full since it compiles from source)
[19:36:27] <davidjames> benchan, adlr: debug symbols for x86-alex-binary should work fine for x86 pfq
[19:36:31] <adlr> how can i make the x86 pfq include symbols?
[19:36:43] <davidjames> adlr: One line change to cbuildbot_config.py :)
[19:36:57] <adlr> ooooo. sounds likea  good idea, no?
[19:37:05] <davidjames> adlr: Sure I'll send you a patch
[19:38:39] <davidjames> adlr: If you like I can also make the crashes symbolized in same patch
[19:39:23] <adlr> davidjames: yes yes yes!
[19:46:30] <davidjames> adlr: Ok, well, here's the first part about getting the symbols: http://gerrit.chromium.org/gerrit/7333
[19:46:42] <adlr> looking
[19:51:41] <benchan> I tried the x86-generic full and x86-alex full binary but couldn't get a fully symbolized stack trace with the crash dmp in the VMtest
[19:59:21] <davidjames> benchan: Well x86-generic full doesn't have a crash, and the symbols coming from that bot don't apply to any other bot
[19:59:37] <davidjames> benchan: x86-alex binary have symbols that also work with PFQ
[20:00:08] <davidjames> benchan: zgb binary has a nice crash on it
[20:01:24] <benchan> davidjames: thanks, I'm downloading from another bot
[20:02:29] <davidjames> benchan: Generating symbols on zgb works for me fine and gets a juicy stacktrace
[20:02:48] *** saintlou|away is now known as saintlou
[20:03:44] <nirnimesh> howdy folks. I'm looking at 20109. Do we have symbolicated stack trace for the other chrome crashes?
[20:03:58] *** rbyers_ has quit IRC
[20:05:08] <davidjames> nirnimesh: Yeah zgb symbolizes fine
[20:05:26] <davidjames> nirnimesh: Also x86-generic-full has some errors about pyauto you might be interested in
[20:05:33] <davidjames> nirnimesh: Unhandled JSONInterfaceError: Automation call {'username': 'performance.test.account at googleapps dot com', 'password': 'perfsmurf', 'command': 'Login'} received empty response.  Perhaps the browser crashed.
[20:05:37] <nirnimesh> Where is it? (sorry I'm late to the party)
[20:06:32] <ellyjones> are we back to trybots?
[20:07:45] <davidjames> nirnimesh: http://www.corp.google.com/~davidjames/zgb-stack.txt
[20:10:02] <nirnimesh> davidjames: this is a crash quite early in the browser startup. Time to loop in the ui folks
[20:10:28] *** Solet has quit IRC
[20:13:52] <benchan> davidjames: thanks for the stack trace.  ar, it's a callback :(
[20:14:27] <davidjames> benchan: Yeah, gee, really hard to figure out what's actually happening there
[20:15:01] <nirnimesh> It may or may not be related to enabling webui login yesterday. hard to tell
[20:16:05] <nirnimesh> BTW, how do you get the symbolicated stack traces? Is there a doc I'm missing?
[20:16:40] <davidjames> adlr is updating the docs
[20:16:51] <davidjames> nirnimesh: But for now you can see http://code.google.com/p/chromium-os/issues/detail?id=17898#c29
[20:17:07] <adlr> vpalatin: benchan: have you looked at http://build.chromium.org/p/chromiumos/builders/x86%20generic%20PFQ/builds/733/steps/VMTest/logs/stdio ?
[20:17:31] <davidjames> adlr: Next run of PFQ should include debug.tgz btw
[20:17:38] <davidjames> adlr: Current one will miss it because it missed my commit
[20:17:46] <adlr> i wonder if that is a chrome crash, or the new web ui doesn't handle that automation call
[20:17:56] <vpalatin> adlr: you mean the CryptoHomeUnmounted issue ?
[20:18:01] <adlr> vpalatin: yes
[20:18:32] <nirnimesh> adlr: no, I had tried that test with webui login before. It had worked
[20:18:40] <adlr> i see
[20:18:44] <benchan> adlr: not yet, i'm still on CryptoHomeMounted
[20:19:21] <vpalatin> adlr: I'm downloading the artifacts to try to run it locally. but I have not much ideas about what to do...
[20:19:30] <adlr> i see
[20:20:10] * eggy blinks
[20:21:50] *** PuffTheMagic has joined #chromium-os
[20:22:34] <PuffTheMagic> does anyone in here know how to get a boot console over ttyHSL0?
[20:22:41] <PuffTheMagic> is that available over usb?
[20:22:50] <PuffTheMagic> is there anyway to get bootconsole over USB?
[20:22:57] *** Solet has joined #chromium-os
[20:25:50] <vpalatin> PuffTheMagic: did you try using console= argument on the kernel command line ?
[20:26:42] <PuffTheMagic> yes, I just cant find much info about ttyHSL0 and that seems to be the default on many HTC msm devices
[20:26:59] <PuffTheMagic> and when i google i see it mentioned in the chromium-os commits
[20:27:10] <PuffTheMagic> so i figured someone around here might have more info about it
[20:27:43] <benchan> adlr, davidjames, vpalatin: i'm wondering why the vmtest crash only happened in x86 builds.
[20:28:15] <davidjames> benchan: arm builds don't run tests
[20:28:21] <vpalatin> benchan: because we have no BVT on ARM ...
[20:28:28] <benchan> davidjames: oh
[20:30:05] <nirnimesh> Is anyone able to repro the crashes locally (on vm or device)?
[20:30:34] *** Styx has quit IRC
[20:31:34] * adlr goes to the long meeting
[20:32:28] <adlr> i'll be back ~1pm
[20:32:54] <vpalatin> adlr: good luck , we are keeping some bugs hot for you
[20:33:47] *** rbyers_ has joined #chromium-os
[20:34:56] <seanpaul> PuffTheMagic: http://htc-linux.org/wiki/index.php?title=Msm_Serial useful?
[20:39:21] <benchan> vpalatin, davidjames: another thing I notice from the history of x86-zgb is that VMTest used to pass with the same version of chrome yesterday
[20:40:01] <vpalatin> benchan: yes chrome version is pinned to 871
[20:40:23] <vpalatin> benchan: you mean the regression is not in chrome ?
[20:40:48] <davidjames> benchan: You are right, that is a really good point
[20:41:28] <benchan> vpalatin: i'm wondering if some change in a package, which chrome depends on, causes the crash
[20:41:51] <davidjames> benchan: It was working at 6pm yesterday, so something must have broken after that time
[20:42:23] <benchan> vpalatin, davidjames: i'm doing a diff on the build logs and trying to narrow down the changes
[20:43:37] *** petermayo has joined #chromium-os
[20:43:37] *** ChanServ sets mode: +v petermayo
[20:44:06] <davidjames> benchan: The hard part about diagnosing it is that test suite was already flaky and just got worse
[20:44:30] <davidjames> From somewhat flaky to failing consistently
[20:45:38] *** rbyers_ has quit IRC
[20:45:46] 
[20:48:09] <vpalatin> davidjames, benchan, adlr: with downloaded artifacts, I can reproduce the login_CryptohomeUnmounted failure
[20:49:38] <sosa> vpalatin:  can you disable webui login on that device by changing session_manager_setup and see if you can still repro?
[20:50:38] <vpalatin> sosa: ok I will try that
[20:52:14] *** alicet has joined #chromium-os
[20:52:14] *** ChanServ sets mode: +v alicet
[20:53:21] <vpalatin> sosa: just removing --webui-login is ok ?
[20:53:55] <sosa> yeah looking at the CL that should be fine
[20:54:05] <sosa> the rest of (http://gerrit.chromium.org/gerrit/#patch,sidebyside,7273,1,session_manager_setup.sh) is just refactoring
[20:54:59] <ellyjones> 421571  localtime=Sep 07 14:52:51       Unhandled PackageInstallError: Installation of pyauto_dep(type:dep) failed : dep-pyauto_dep.tar.bz2 could not be fetched from any of the repos ['autoserv://']
[20:55:04] <ellyjones> anyone seen that one around?
[20:55:22] <nirnimesh> ellyjones: --use_emerged
[20:55:40] <ellyjones> why does it not work by default?
[20:56:10] <nirnimesh> because pyauto_dep is built by the chromeos-chrome ebuild, which writes to /build/$BOARD only, not to your workspace
[20:57:38] <ellyjones> ahh
[20:58:50] <crosbot> tree became '?'
[20:59:05] <crosbot> tree became 'Tree is closed (need to sort out crosbug.com/20109 and other VMtest crashes)'
[20:59:56] <sosa> nirnimesh: does that mean run_remote_tests --build will never work with smoke_suite?
[21:01:35] <nirnimesh> sosa: what does --build do? It should work with --use_emerged
[21:01:57] <sosa>   if [ ${FLAGS_use_emerged} -eq ${FLAGS_TRUE} ]; then
[21:01:57] <sosa>     AUTOTEST_DIR="/build/${FLAGS_board}/usr/local/autotest"
[21:01:57] <sosa>     FLAGS_build=${FLAGS_FALSE}
[21:03:16] <sosa> ah nvm, it's probably fine...but --build let's you easily develop on tests with run_remote_tests as it copies just changes in the autotest tests to the running autotest dir
[21:03:31] <sosa> so if you're making simple changes to tests etc, you don't have to re-emerge all the autotest items
[21:03:45] <sosa> meaning you don't have to wait 2-3 minutes between tests...rather about 5 seconds
[21:04:12] <crosbot> tree became '?'
[21:04:28] <crosbot> tree became 'Tree is closed (need to sort out crosbug.com/20109 and other VMtest crashes)'
[21:04:42] <nirnimesh> sosa: yes, I understand that was easy, but zbehan tells me that --use_emerged should become the default. When making test edits, I manually cp to /build/$BOARD/usr/local/autotest/..
[21:05:17] <sosa> that is not a very good workflow
[21:05:28] *** rbyers_ has joined #chromium-os
[21:05:29] *** ChanServ sets mode: +v rbyers_
[21:05:56] <ellyjones> login_CryptohomeIncognitoUnmounted failed for me oO
[21:05:58] <sosa> regardless of what zbehan says, many many devs use --build with run_remote_tests including those who wrote it :p
[21:06:09] <sosa> without --build is just too slow.
[21:06:19] <ellyjones> yes, re-emerging autotests is terribly slow
[21:06:42] <ellyjones> check _this_ out:
[21:06:43] <ellyjones> 15:04:14 INFO |                 ERROR   login_CryptohomeIncognitoUnmounted      login_CryptohomeIncognitoUnmounted   timestamp=1315422253     localtime=Sep 07 15:04:13       Error during login (chronos at gmail dot com, chronos): Google authentication failed..
[21:06:56] <ellyjones> why is CryptohomeIncognitoUnmounted during authentication at all?
[21:07:54] <nirnimesh> sosa: I don't know the ebuild system well enough to give an authoratative answer. All I know is that you'd have to use --use_emerged for stuff that gets built by the chromeos-chrome ebuild
[21:09:40] <sosa> sounds like we can fix that :)
[21:09:50] <sosa> rather than a "must have"
[21:12:27] <ellyjones> oh argh
[21:12:28] <ellyjones> closed tree
[21:13:21] *** FX80 has quit IRC
[21:15:36] <davidjames> hey, tests passed on pfq!
[21:15:43] *** dgarrett has joined #chromium-os
[21:15:44] *** ChanServ sets mode: +v dgarrett
[21:16:06] <davidjames> other builders are looking good as well, did someone fix the tree? :)
[21:16:22] *** alicet has quit IRC
[21:16:32] <crosbot> tree became '?'
[21:16:47] <crosbot> tree became 'Tree is closed (need to sort out crosbug.com/20109 and other VMtest crashes)'
[21:17:24] <vpalatin> davidjames: sadly, I don't see many relevant commits between the failed PFQ and the successful one
[21:18:27] <davidjames> vpalatin: Latest TOT preflight queue is interesting, desktopui_WindowManagerFocusNewWindows crashed Chrome
[21:20:50] <benchan> davidjames, vpalatin: it may be useful to show the packages being uprev in the log
[21:21:04] <davidjames> benchan: They are shown in the log
[21:24:12] *** jshin has joined #chromium-os
[21:24:12] *** ChanServ sets mode: +v jshin
[21:25:19] *** rbyers_ has quit IRC
[21:25:25] <benchan> davidjames: I mean in a highlighted form
[21:25:46] *** SoulRaven has quit IRC
[21:28:08] <vpalatin> sosa, davidjames, adlr, benchan: here is the result : with the same artifacts binaries from a failed PFQ, I get a login_CryptohomeUnmounted failure  on KVM without touching anything and a success by removing --webui-login
[21:28:29] <davidjames> vpalatin: Thanks can you remove --webui-login then?
[21:28:57] <vpalatin> davidjames: you mean reverting http://gerrit.chromium.org/gerrit/#change,7273 ?
[21:29:30] <davidjames> vpalatin: Yes
[21:30:08] *** rbyers_ has joined #chromium-os
[21:30:08] *** ChanServ sets mode: +v rbyers_
[21:30:10] <davidjames> vpalatin: Since that'll open the tree
[21:31:19] <gauravsh> ellyjones: That login_CryptohomeIncognitoUnmounted thing seems to be wrong. I am not sure why the control file specifies a user credential to use.
[21:31:21] <benchan> vpalatin, davidjames: is the recent pass in VMTest in PFQ due to that change?
[21:31:37] <ellyjones> I mailed ericli to ask about it
[21:32:01] <gauravsh> ellyjones: cmasone wrote that - not sure why he choose to use $backdoor credentials in the control file for the test.
[21:32:38] <ellyjones> oh, okay
[21:33:03] <gauravsh> since he is out galivanting around the country, guess we can just wait till he gets back.
[21:33:39] <vpalatin> davidjames, benchan: http://gerrit.chromium.org/gerrit/#change,7347
[21:34:14] <davidjames> vpalatin: thanks, approved
[21:34:15] <crosbot> tree became '?'
[21:34:31] <crosbot> tree became 'Tree is closed (need to sort out crosbug.com/20109 and other VMtest crashes)'
[21:36:01] <benchan> vpalatin: thanks. so the recent PFQ pass is due to flakiness?
[21:37:01] <vpalatin> benchan: probably
[21:39:03] <benchan> vpalatin: there is a merge conflict
[21:40:05] <vpalatin> benchan: yes I'm working on it
[21:40:26] <vpalatin> the "revert" button is magic excepted when there is a merge conflict afterwards ...
[21:43:36] <davidjames> benchan: Yeah I think it's flakiness, tot preflight queue failed again so it looks like the failures are off and on
[21:45:09] <benchan> davidjames: let's see how it goes with the revert
[21:46:44] <ellyjones> "Autotest client terminated unexpectedly" = ?
[22:06:11] <PuffTheMagic> seanpaul, yes thanks
[22:08:25] <ttuttle|work> okay, I give up
[22:08:34] <ttuttle|work> I don't have /sys/kernel/debug/preserved, and I don't have any idea why.
[22:08:50] <ttuttle|work> anyone have any ideas?
[22:09:01] <ttuttle|work> (I want the kernel log from last boot, as I'm debugging a panic.)
[22:10:53] <vpalatin> benchan: heading to lunch
[22:12:45] *** tbroch has quit IRC
[22:23:22] *** rbyers_ has quit IRC
[22:26:06] <davidjames> benchan: Ran a PFQ locally with the revert, it passed fine
[22:26:54] <ellyjones> woo
[22:27:06] <ellyjones> does that mean I can land a risky cl now?
[22:27:42] <davidjames> ellyjones: Did you test it with the trybot?
[22:27:58] <davidjames> ellyjones: If it's a risky CL, probably worth running through trybot on a few platforms first
[22:28:51] <ellyjones> how do I do that?
[22:29:14] <davidjames> ellyjones: http://www.chromium.org/chromium-os/local-trybot-documentation
[22:29:36] <davidjames> run cbuildbot --list to see a list of common configs to run with.  Use the --all flag to see all possible configs.
[22:30:18] <ellyjones> cool
[22:31:08] *** saintlou has quit IRC
[22:31:13] <davidjames> vpalatin, benchan, adlr: Looks like PFQ passed too, time to open?
[22:31:52] <nirnimesh> what's the resolution? The crashes were due to the webui login?
[22:32:04] *** saintlou has joined #chromium-os
[22:32:04] *** ChanServ sets mode: +v saintlou
[22:34:07] <vpalatin> davidjames: yes we will
[22:34:46] <adlr> i think we are rolling back webui login
[22:36:12] <benchan> vpalatin, adlr, davidjames: the recent x86 PFQ is green, do we need to wait for another green to reopen the tree?
[22:36:14] <adlr> * vpalatin is doing the honors
[22:36:19] <crosbot> tree became 'Tree is open (webui login activation reverted :('
[22:36:36] <benchan> nevermind, it's open :)
[22:36:41] <davidjames> benchan: Probably OK to open now... if it fails again then re-close
[22:36:49] <adlr> now we need to see if chrome is still pinned
[22:36:55] <adlr> we should probably un-pin it
[22:38:24] <davidjames> adlr: If you want to unpin Chrome you'll probably want to run a Chrome PFQ run first
[22:38:30] <davidjames> adlr: If that passes then you can unpin Chrome
[22:38:54] <adlr> i see
[22:39:18] <vpalatin> adlr: we need to get everything cycling green first ?
[22:39:20] <adlr> so does pinning not affect x86 generic TOT chrome PFQ?
[22:39:41] <davidjames> adlr: It doesn't affect either of the Chrome PFQs
[22:39:47] <adlr> ah, okay
[22:39:51] <davidjames> adlr: Chrome PFQs unpin chrome using package.keywords
[22:39:57] <sosa> yeah...if the main Chrome PFQ is ok with revving it with these reverts
[22:40:01] <vpalatin> adlr, benchan: I have clobbered the x86-generic-full as it was going to fail
[22:40:08] <davidjames> adlr: So if you pin chrome using that method, the Chrome PFQs can unpin
[22:40:09] <sosa> then we should look at unpinning
[22:40:21] <adlr> sosa: vpalatin: should we clobber "x86 generic TOT chrome PFQ"?
[22:40:24] <davidjames> adlr: But yeah I would kick the regular Chrome PFQ and see if it passes now
[22:40:31] <davidjames> adlr: Not TOT chrome, regular Chrome PFQ
[22:40:35] <adlr> oh
[22:40:36] <sosa> adlr: ^^^ what davidjames said
[22:40:47] <adlr> okay, i'll clobber "x86 generic chrome PFQ"?
[22:40:57] <adlr> if taht passes, i'll unpin chrome
[22:41:02] <sosa> because the non-TOT pfq is actually what revs the version we build -- it builds based on DEPS from chrome rather than svn revions.  yup
[22:41:25] <adlr> vpalatin: benchan: LGTY?
[22:41:33] <davidjames> adlr: Exactly, but you don't need to clobber, just a normal force build
[22:41:41] <adlr> oh okay
[22:41:42] <sosa> adlr:  shouldn't need to "clobber".  a force build should be  be fine.
[22:41:45] <sosa> yeah
[22:41:51] <adlr> okay, doing that now
[22:41:55] <sosa> adlr: also you need to wait for the pfq
[22:42:01] <adlr> which pfq?
[22:42:04] <sosa> adlr:  to finish pushing the rev for session_manager
[22:42:18] <sosa> adlr: the main one that just went green
[22:42:23] <adlr> oh, so wait for "x86 generic PFQ"?
[22:42:31] <sosa> adlr:  it needs to publish what it knows to the other builders :)
[22:42:32] <davidjames> adlr: Ah, but I think it just finished and pushed the session manager change no?
[22:42:56] <sosa> adlr,davidjames: oh wait you're right ... i'm slow ... i had a lasttime= set
[22:43:14] <sosa> force build away!
[22:43:19] <adlr> okay, force rebuilding on "x86 generic chrome PFQ"
[22:43:42] <adlr> forced!
[22:46:14] *** kinaba_ has joined #chromium-os
[22:46:15] *** ChanServ sets mode: +v kinaba_
[22:46:22] *** kinaba_ has quit IRC
[22:46:40] *** kinaba_ has joined #chromium-os
[22:46:40] *** ChanServ sets mode: +v kinaba_
[22:47:21] *** saintlou has quit IRC
[22:51:10] *** arun_ has quit IRC
[22:58:23] *** saintlou has joined #chromium-os
[22:58:23] *** ChanServ sets mode: +v saintlou
[23:04:31] <nirnimesh> the artifacts doesn't contain the autotest package. how do you guys run the tests on it?
[23:06:38] <adlr> i think it does for some trees
[23:09:14] <adlr> nirnimesh: where are you not seeing it?
[23:09:40] <nirnimesh> I'm looking at the artifacts (followed the buildbot links to Google Storage)
[23:10:03] <sosa> yeah we don't add them in for PFQ's because it takes a while to generate
[23:10:20] <sosa> and it's pretty easy to sync TOT and run autotest for pretty much the same effect
[23:16:21] <benchan> adlr: looks like x86 generic chrome PFQ turns green
[23:16:28] * adlr looks
[23:16:42] <adlr> what's the purple about?
[23:16:56] <adlr> "offline
[23:16:58] <adlr> 1 pending"?
[23:17:16] <benchan> adlr: weird, the bot is running
[23:17:47] <adlr> okay, i will unpin chrome and send this for review
[23:20:43] <crosbot> tree became 'Tree is closed (Automatic: "cbuildbot_master" on "TOT Pre-Flight Queue" from a8ca1a99ddee3b8c1c60bddb27c0c6b4893ec92a: finnur at chromium dot org <finnur at chromium dot org@0039d316-1c4b-4281-b951-d872f2087c98>)'
[23:21:28] <nirnimesh> vpalatin: did you say you were able to repro the webui failures on kvm?
[23:21:48] <nirnimesh> can you tell me which image you tried on?
[23:22:10] <adlr> which tree is "TOT Pre-Flight Queue"?
[23:22:36] <vpalatin> adlr: on the internal waterfall
[23:22:42] <vpalatin> (the first one from left)
[23:22:42] <benchan> vpalatin: TOT PFQ caught a sig 11 on login_BadAuthentication, seems like crosbug.com/20109
[23:23:21] <vpalatin> benchan: yes this has started before the rollback AFAIK
[23:23:50] <vpalatin> hummmm not sure
[23:24:01] <adlr> i think we can open
[23:24:03] <adlr> it was before the rollback
[23:24:15] <benchan> vpalatin: i'm downloading the dump file to take a look
[23:24:56] <vpalatin> nirnimesh: you can use that one I think : https://sandbox.google.com/storage/chromeos-image-archive/x86-generic-pre-flight-queue/0.15.1005.0-a1-b735
[23:25:32] <adlr> sosa: others: to unpin chrome, is it enough to just comment out the line in package.keywords?
[23:25:58] <vpalatin> adlr: the rollback timestamp seems to be 12:45 and this build starts at 13:19 ?
[23:26:39] <sosa> adlr: you should also delete the old chrome ebuild for cleanup.
[23:26:46] <adlr> i see your rollback at 1:23pm, but the failed build started at 1:19pm
[23:27:04] <adlr> sosa: okay, thanks. i assume comment out the line first, then delete ebuild
[23:27:38] <sosa> adlr:  you can double check and emerge chromeos-chrome afterwards as a sanity check.  see that it's about to do the right thing
[23:27:50] <adlr> okay
[23:28:14] <vpalatin> adlr: yes it's 13:23 on the internal (despite being 12:45 on the external)
[23:28:21] <vpalatin> thus I will reopen
[23:28:57] <crosbot> tree became 'Tree is open (retrying after webui rollback)'
[23:30:59] *** saintlou has quit IRC
[23:31:31] *** nkostylev has quit IRC
[23:31:31] *** glotov has quit IRC
[23:33:41] <adlr> CL to unpin chrome: http://gerrit.chromium.org/gerrit/#change,7353
[23:34:07] *** nkostylev has joined #chromium-os
[23:34:08] *** ChanServ sets mode: +v nkostylev
[23:34:16] *** glotov has joined #chromium-os
[23:34:17] *** ChanServ sets mode: +v glotov
[23:34:36] <davidjames> vpalatin, adlr: The TOT chrome build that failed did pick up the webui revert
[23:34:56] <davidjames> vpalatin: You can see that under LKGMCandidateSYnc
[23:35:04] <davidjames> vpalatin: It mentions [vpalatin:7347]
[23:35:24] <davidjames> vpalatin: Oh, but PFQ hadn't rev'd it yet, of course
[23:35:48] <sosa> davidjames:  yeah it's a bit wrong with workon and the chrome pfq's
[23:36:15] <sosa> davidjames:  i should  disable the lkgm sync blamelist on the Chrome PFQ's
[23:36:58] <davidjames> sosa: That was for the internal bot
[23:37:03] <davidjames> sosa: That was for the internal TOT
[23:37:18] <davidjames> (oh, yeah I said TOT chrome, meant Internal TOT PFQ)
[23:37:39] <davidjames> sosa: But yeah, your comment applies there as well
[23:40:12] <adlr> chrome has been unpinned
[23:40:34] <davidjames> adlr: Woohoo!
[23:41:37] * adlr documents unpinning chrome in theFAQ
[23:48:27] <marcheu> vpalatin: for what it's worth, the amd64-generic build was never green, but thanks for working on it :)
[23:49:56] *** behdad has quit IRC
[23:56:58] <nirnimesh> xiyuan and I tried the tests on a several images. Cannot repro the crashes :(





top