Fix up the SSHHost.__del__ method to use the set.discard instead of
set.remove when removing the warning logger stream from the current
set of warning loggers to be listened on, since the stream may have
already been removed for other reasons (e.g. the logging process
terminated and job.record already noticed this and removed the
stream from the set).
Signed-off-by: John Admanski <jadmanski@google.com>
git-svn-id: http://test.kernel.org/svn/autotest/trunk@1266 592f7852-d20e-0410-864c-8624ca9c26a4
diff --git a/server/hosts/ssh_host.py b/server/hosts/ssh_host.py
index 977674c..a3ea3ca 100644
--- a/server/hosts/ssh_host.py
+++ b/server/hosts/ssh_host.py
@@ -114,7 +114,7 @@
# kill the console logger
if getattr(self, 'logger_popen', None):
self.__kill(self.logger_popen)
- self.job.warning_loggers.remove(self.warning_stream)
+ self.job.warning_loggers.discard(self.warning_stream)
self.warning_stream.close()
# kill the netconsole logger
if getattr(self, 'netlogger_popen', None):