[checkmk-commits] #3037 FIX Terminate datasource programs when check timeout occurs

Module: check_mk Branch: master Commit: 3240e2270ea0d7c0ada63f648a79a6180e0e3d26 URL: http://git.mathias-kettner.de/git/?p=check_mk.git;a=commit;h=3240e2270ea0d7… Author: Lars Michelsen <lm(a)mathias-kettner.de> Date: Fri Jan 29 13:23:22 2016 +0100 #3037 FIX Terminate datasource programs when check timeout occurs When using datasource programs, e.g. for checking by SSH, the Check_MK process could be terminated by a reached timeout leaving the launched subprocesses open. This might have caused in processes being too long open, running and consuming resources. The datasource program process and it's sub processes are now cleaned up when a check timeout occures. --- .werks/3037 | 15 +++++++++++++++ ChangeLog | 1 + modules/check_mk_base.py | 6 +++++- 3 files changed, 21 insertions(+), 1 deletion(-) diff --git a/.werks/3037 b/.werks/3037 new file mode 100644 index 0000000..e843806 --- /dev/null +++ b/.werks/3037 @@ -0,0 +1,15 @@ +Title: Terminate datasource programs when check timeout occurs +Level: 2 +Component: core +Compatible: compat +Version: 1.2.7i4 +Date: 1454069860 +Class: fix + +When using datasource programs, e.g. for checking by SSH, the Check_MK +process could be terminated by a reached timeout leaving the launched +subprocesses open. This might have caused in processes being too long +open, running and consuming resources. + +The datasource program process and it's sub processes are now cleaned +up when a check timeout occures. diff --git a/ChangeLog b/ChangeLog index ecd02aa..87e13fa 100644 --- a/ChangeLog +++ b/ChangeLog @@ -14,6 +14,7 @@ * 2894 FIX: Fixed broken cluster checks when using Nagios core... * 2926 FIX: Fix exception in WATO when a check man page has not catalog entry * 2944 FIX: Fixed exception during config activation having clusters with unresolvable node IP addresses + * 3037 FIX: Terminate datasource programs when check timeout occurs... Checks & Agents: * 2434 NetApp monitoring: Cluster-Mode is now supported, changes in existing 7Mode checks... diff --git a/modules/check_mk_base.py b/modules/check_mk_base.py index 995e457..eef1b9c 100644 --- a/modules/check_mk_base.py +++ b/modules/check_mk_base.py @@ -772,10 +772,14 @@ def get_agent_info_program(commandline): vverbose("Calling external program %s\n" % commandline) try: - p = subprocess.Popen(commandline, shell = True, stdout = subprocess.PIPE, stderr = subprocess.PIPE) + p = subprocess.Popen(commandline, shell = True, stdout = subprocess.PIPE, + stderr = subprocess.PIPE, preexec_fn=os.setsid) stdout, stderr = p.communicate() exitstatus = p.returncode except MKTimeout: + # On timeout exception try to stop the process to prevent child process "leakage" + if p: + os.killpg(os.getpgid(p.pid), signal.SIGTERM) raise except Exception, e: