Issue metadata
Sign in to add a comment
|
webkit_tests failing on 2 builders: wptserve.py issue in _stop_running_server |
||||||||||||||||||||||
Issue descriptionwebkit_tests failing on 2 builders Type: build-failure Builders failed on: - WebKit Linux Trusty (dbg): https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29 - WebKit Linux Trusty MSAN: https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20MSAN """ 22:31:46.542 11072 Found 50839 tests; running 41935, skipping 8904. 22:31:46.542 11072 Checking build ... 22:31:46.610 11072 22:31:46.778 11072 "ruby --version" took 0.17s 22:31:46.823 11072 "/usr/bin/wdiff --help" took 0.03s 22:31:47.006 11072 "/usr/sbin/apache2 -v" took 0.18s 22:31:47.007 11072 Checking system dependencies ... 22:31:48.160 11072 "/b/c/b/linux_layout/src/out/Release/content_shell --check-layout-test-sys-deps" took 1.07s 22:31:48.161 11072 Clobbering old results in /b/rr/tmpIQ7wPg/w 22:31:48.194 11072 Starting WPTServe ... 22:31:48.207 11072 stale wptserve pid file, pid 13459 22:31:48.207 11072 Flushing stdout 22:31:48.207 11072 Flushing stderr 22:31:48.207 11072 Cleaning up port AttributeError raised: 'NoneType' object has no attribute 'wait' Traceback (most recent call last): File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/run_webkit_tests.py", line 68, in main return run(port, options, args, stderr, stdout).exit_code File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/run_webkit_tests.py", line 565, in run run_details = _run_tests(port, options, args, printer) File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/run_webkit_tests.py", line 556, in _run_tests return manager.run(args) File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/controllers/manager.py", line 127, in run self._start_servers(tests_to_run) File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/controllers/manager.py", line 365, in _start_servers self._port.start_wptserve() File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/port/base.py", line 1248, in start_wptserve server.start() File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/servers/server_base.py", line 93, in start self._stop_running_server() File "/b/c/b/linux_layout/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/servers/wptserve.py", line 76, in _stop_running_server self._process.wait() AttributeError: 'NoneType' object has no attribute 'wait' <Thread(Thread-1, started 140542164846336)> ProcessRead: proc.stdout finished. <Thread(Thread-1, started 140542164846336)> ProcessRead: cleaning up. <Thread(Thread-2, started daemon 140542156453632)> TimedFlush: Finished <Thread(Thread-1, started 140542164846336)> ProcessRead: finished. Confused: 8 files were deleted from /tmp during the test run Stopping Xvfb with pid 10908 ... Xvfb pid file removed killed dbus-daemon with PID 10891 cleared DBUS_SESSION_BUS_ADDRESS environment variable step returned non-zero exit code: 255 @@@STEP_FAILURE@@@ """
,
Nov 16 2016
,
Nov 16 2016
So, at some point the attribute _process is set to None when it's expected to be set; it's supposed to be set at https://cs.chromium.org/chromium/src/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/servers/server_base.py?l=165 in _spawn_process, but if self._executive.popen fails then it could perhaps be set to None. I can't think of any obvious recent changes that I think would cause this; the root cause of this issue may not be new, but it may be manifesting now that we're using wpsterve by default.
,
Nov 16 2016
Also, note: wptserve is being started and stopped successfully on the latest builds for the builders - https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29 - https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20MSAN Note that both of these builders are new; they're newly added Trusty builders that are replacing Precise builders ( issue 660580 ), so maybe it was a temporary problem that manifested in the first few builds?
,
Nov 16 2016
These two bots look to have no such exception now. Let's close this, and reopen if the problem happens again.
,
Nov 17 2016
tkent@ - This error was most certainly hiding what the real failure was. There are multiple ways this could have occurred, so can we please fix this code so that it doesn't fail if self._process is None?
,
Nov 17 2016
Okay, I created https://codereview.chromium.org/2511013002 to fix this issue.
,
Nov 17 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/8a66268ab69e0374e1974e70fdfbe1a51980699c commit 8a66268ab69e0374e1974e70fdfbe1a51980699c Author: tansell <tansell@chromium.org> Date: Thu Nov 17 04:59:52 2016 webkitpy: Refactor _stop_running_server for wptserve * Always make sure the process has terminated. * Always clean up the pid file if it exists. * Don't fail if process was never started. BUG= 665769 R=tkent,qyearsley Review-Url: https://codereview.chromium.org/2511013002 Cr-Commit-Position: refs/heads/master@{#432750} [modify] https://crrev.com/8a66268ab69e0374e1974e70fdfbe1a51980699c/third_party/WebKit/Tools/Scripts/webkitpy/layout_tests/servers/wptserve.py
,
Nov 18 2016
This is still failing consistently 24 hours later. Did your fix not work?
,
Nov 18 2016
mgiuca@, what build# failed with this issue? I couldn't find in https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29
,
Nov 18 2016
Sorry, I didn't properly read the issue. I thought this was related to the failure of network-disable-cache-preloads.php and unitless-length.html (it was incorrectly linked in Sheriff-O-Matic). |
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by suzyh@chromium.org
, Nov 16 2016Status: Assigned (was: Available)