maryjane
/
syncevolution
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

D-Bus testing: kill partially terminated processes 

When running syncevo-dbus-server under valgrindcheck.sh,
the following happened occasionally:
- syncevo-dbus-server main thread quits, some threads keep running
  => ps shows the process as <defunct> with ppid = 1 = init
- valgrindcheck.sh notices that the process is done,
  reports status and quits
- test-dbus.py fails to wait for the syncevo-dbus-server
  process (because it is not the parent) and assumes that
  the process is gone

At this point there is a lingering process which occupies the
well-known D-Bus name (= all further tests fail) and which
prevents unmounting the chroot.

It's unknown how the syncevo-dbus-server gets into that state.
Could be valgrind 3.7.0 or the kernel 3.4.28-2.20-xen.

As a workaround, let test-dbus.py collect the pids of all processed that it
couldn't wait for and send them SIGKILLs until that returns with "not
found".
This commit is contained in:
Patrick Ohly 2013-03-06 01:59:11 -08:00
parent 762d6c0910
commit ef4f77e06a
1 changed files with 35 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
test/test-dbus.py
View File

@ -786,10 +786,27 @@ Use check=lambda: (expr1, expr2, ...) when more than one check is needed.
logging.printf("found children: %s", children)
return children
def killPending(self, pending):
'''Ensure that all processes listed with their pid are not running.'''
while True:
again = []
for pid in pending:
logging.printf("sending SIGKILL to pending proccess %d" % pid)
if not TryKill(pid, signal.SIGKILL):
logging.printf("pending process %d is gone" % pid)
else:
again.append(pid)
if again:
time.sleep(0.1)
pending = again
else:
break
def killChildren(self, delay):
'''Find all children of the current process and kill them. First send SIGTERM,
then after a grace period SIGKILL.'''
children = self.getChildren()
pending = []
# First pass with SIGTERM?
if delay:
for pid, (name, cmdline) in children.iteritems():
@ -819,9 +836,24 @@ Use check=lambda: (expr1, expr2, ...) when more than one check is needed.
del children[pid]
except OSError, ex:
if ex.errno == errno.ECHILD:
# someone else must have been faster, also okay
# Process might have transferred to init as parent:
#
# PID PPID TID NLWP CMD COMMAND
# 20617 1 - 2 [memcheck-amd64-] <defunct> [memcheck-amd64-] <defunct>
# - - 20617 - - -
# - - 21012 - - -
#
# This was observed for syncevo-dbus-server when running under
# valgrind despite apparently having terminated normally
# (valgrindcheck.sh sees the exist status, log output confirms
# normal shutdown). Not sure why process shutdown did not
# terminate all threads.
#
# Remember to kill the remaining threads with SIGKILL,
# as a workaround.
logging.printf("process %d %s gone at %s",
pid, name, time.asctime())
pending.append(pid)
del children[pid]
else:
raise ex
@ -829,6 +861,7 @@ Use check=lambda: (expr1, expr2, ...) when more than one check is needed.
# All children quit normally.
logging.printf("all process gone at %s",
time.asctime())
self.killPending(pending)
return []
time.sleep(0.1)
# Force killing of remaining children. It's still possible
@ -841,6 +874,7 @@ Use check=lambda: (expr1, expr2, ...) when more than one check is needed.
if TryKill(pid, signal.SIGKILL):
logging.printf("killed %d %s", pid, name)
killed.append("%d %s" % (pid, name))
self.killPending(pending)
return killed
def serverExecutableHelper(self, pid):