cfagent dying suddenly.

[Hans Deragon]

Greetings.


  I noticed that once in a while, cfagent would die unexpectedly.  It happens 
with both versions 2.0.7p3 and 2.1.0a6.  It only occurs when I run numerous cf 
scripts at once.  It happens not very often though, maybe once every 20 runs.

  I have a hierarchy of cf scripts.  My first cf script imports other cf 
scripts, which themselves import other cf scripts until ¨leaf¨ cf scripts are 
called and executed.  If I run ¨leaf¨ cf scripts directly, cfagent never dies.

  cfagent never leaves a core file or reports a signal.  It just stops, as if a 
clean exit(0) call was made (I checked $? just after it died and it contained 0).

  Anybody experienced the same problem or have some comments?  At the end of 
this email, you will find the last output gathered before cfagent dies.  It does 
not always die at the same place.


Best regards,
Hans Deragon
--
Consultant en informatique/Software Consultant
Deragon Informatique inc.         Open source:
http://www.deragon.biz            http://swtmvcwrapper.sourceforge.net
mailto://hans@deragon.biz         http://autopoweroff.sourceforge.net



ANNEX:

Following, the output of cfagent just before dying.  It ends with ¨PutLock(¨. 
Notice that does not mean it crashed near PutLock(), but that the stdout output 
probably wasn't flushed when cfagent died.



                  New temporary class additions
                  --------( Pass 1 )-------
                             ssh
Actionsequence item editfiles
Editfile()
WrapDoEditFile(/root/.ssh/authorized_keys2,/root/.ssh/authorized_keys2)
FileObjectFilter(/root/.ssh/authorized_keys2)
DoEditFile(/root/.ssh/authorized_keys2)
IgnoredOrExcluded(/root/.ssh/authorized_keys2)
GetLock(editfile.ssh,_root__ssh_authorized_keys2,time=1063900943), 
ExpireAfter=120, IfElapsed=0
GetLastLock()
CheckOldLock(lock.__lmc_cf.localhost.editfile.ssh._root__ssh_authorized_keys2)
Lock lock.__lmc_cf.localhost.editfile.ssh._root__ssh_authorized_keys2 last ran 
at Thu Sep 18 11:39:21 2003

Couldn't obtain lock for 
lock.__lmc_cf.localhost.editfile.ssh._root__ssh_authorized_keys2 (already running!)
WrapDoEditFile(/etc/ssh/ssh_known_hosts,/etc/ssh/ssh_known_hosts)
FileObjectFilter(/etc/ssh/ssh_known_hosts)
DoEditFile(/etc/ssh/ssh_known_hosts)
IgnoredOrExcluded(/etc/ssh/ssh_known_hosts)
GetLock(editfile.ssh,_etc_ssh_ssh_known_hosts,time=1063900943), ExpireAfter=120, 
IfElapsed=0
GetLastLock()
CheckOldLock(lock.__lmc_cf.localhost.editfile.ssh._etc_ssh_ssh_known_hosts)
Lock lock.__lmc_cf.localhost.editfile.ssh._etc_ssh_ssh_known_hosts last ran at 
Thu Sep 18 09:54:32 2003

cfengine:localhost: Lock 
lock.__lmc_cf.localhost.editfile.ssh._etc_ssh_ssh_known_hosts expired...(after 
127/120 minutes)
Trying to kill expired process, pid 2456
LockLog(Lock expired, process killed)
SetLock(lock.__lmc_cf.localhost.editfile.ssh._etc_ssh_ssh_known_hosts)
PutLock(