https://bugs.webkit.org/show_bug.cgi?id=116355
Reviewed by Stephanie Lewis.
Throttle the number of processes to use, and also bump up the number of crashes/timeouts
we tolerate from 20 to 50.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunWebKitTests.start):
* Scripts/webkitpy/port/mac.py:
(MacPort.default_child_processes):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@150303
268f45cc-cd09-0410-ab3c-
d52691b4dbfc
if platform.startswith('mac'):
self.setCommand(self.command + ['--no-build'])
if shouldAbortEarly:
- self.setCommand(self.command + ["--exit-after-n-crashes-or-timeouts", "20", "--exit-after-n-failures", "500"])
+ self.setCommand(self.command + ["--exit-after-n-crashes-or-timeouts", "50", "--exit-after-n-failures", "500"])
if platform == "win":
rootArgument = ['--root=' + os.path.join("WebKitBuild", self.getProperty('configuration'), "bin")]
self.setCommand(self.command + ['--no-build', '--use-remote-links-to-tests'])
+2013-05-17 Ryosuke Niwa <rniwa@webkit.org>
+
+ Leaks builders exit early frequently
+ https://bugs.webkit.org/show_bug.cgi?id=116355
+
+ Reviewed by Stephanie Lewis.
+
+ Throttle the number of processes to use, and also bump up the number of crashes/timeouts
+ we tolerate from 20 to 50.
+
+ * BuildSlaveSupport/build.webkit.org-config/master.cfg:
+ (RunWebKitTests.start):
+ * Scripts/webkitpy/port/mac.py:
+ (MacPort.default_child_processes):
+
2013-05-17 Peter Gal <galpeter@inf.u-szeged.hu>
Remove chrome-channels command from the webkit-patch
# FIXME: https://bugs.webkit.org/show_bug.cgi?id=95906 With too many WebProcess WK2 tests get stuck in resource contention.
# To alleviate the issue reduce the number of running processes
# Anecdotal evidence suggests that a 4 core/8 core logical machine may run into this, but that a 2 core/4 core logical machine does not.
- if self.get_option('webkit_test_runner') and default_count > 4:
+ should_throttle_for_wk2 = self.get_option('webkit_test_runner') and default_count > 4
+ # We also want to throttle for leaks bots.
+ if should_throttle_for_wk2 or self.get_option('leaks'):
default_count = int(.75 * default_count)
# Make sure we have enough ram to support that many instances: