summaryrefslogtreecommitdiff
path: root/lib/parallel.py
diff options
context:
space:
mode:
authorPrashanth B <beeps@google.com>2014-03-07 13:31:36 -0800
committerchrome-internal-fetch <chrome-internal-fetch@google.com>2014-04-02 23:43:37 +0000
commit74a4c7314f9a639418d1e1074139de1be1d33f9d (patch)
treee5490c8c49635ea03b199bc13943ad13ea53cbe5 /lib/parallel.py
parentfc1e40b7332be25a6d13259b6ee40dae86e4a42e (diff)
downloadchromite-74a4c7314f9a639418d1e1074139de1be1d33f9d.tar.gz
Give canary hwtests time till the next run to complete.
Reasons: 1. we see weird (non fatal) delays in running hwtests 2. devices in pool:bvt are just sitting idle between canaries 3. Tpms ask for timed out suite reruns 4. Build sheriffs are confused about how to handle a timeout 5. Most timeouts happen because of suites competing for DUTs It shouldn't hurt to give the canaries time till the next run. We should find a better way to monitor 1, one that doesn't involve failing a build. TEST=Trybot, unittests. BUG=chromium:342502, chromium:351381 Change-Id: I143fba9b3d2ea6543e1dd3893dd2f1cee813d758 Reviewed-on: https://chromium-review.googlesource.com/192682 Tested-by: Prashanth B <beeps@chromium.org> Reviewed-by: David James <davidjames@chromium.org> Commit-Queue: Prashanth B <beeps@chromium.org>
Diffstat (limited to 'lib/parallel.py')
-rw-r--r--lib/parallel.py2
1 files changed, 1 insertions, 1 deletions
diff --git a/lib/parallel.py b/lib/parallel.py
index b5ef3db7c..49569117e 100644
--- a/lib/parallel.py
+++ b/lib/parallel.py
@@ -76,7 +76,7 @@ class _BackgroundTask(multiprocessing.Process):
# The time we allow processes to be silent. This must be greater than the
# hw_test_timeout set in cbuildbot_config.py, and less than the timeout set
# by buildbot itself (typically, 150 minutes.)
- SILENT_TIMEOUT = 60 * 145
+ SILENT_TIMEOUT = 60 * 205
# The amount by which we reduce the SILENT_TIMEOUT every time we launch
# a subprocess. This helps ensure that children get a chance to enforce the