Nathan Scott (nscott@xxxxxxxxxx) reports a problem with a slightly
modified version of 2.7.4-1, and speculates the same problem could
possibly hit users running a large number of PMDAs on a loaded system.
His problem is fixed by reinstating the longer (20s) delay in the
init.d script. I don't have any issue with queuing this change, but I
don't see getting it out in a release as a priority in the absence of
evidence that it has occurred in the release as distributed.
The symptom is that you see "forcing pmcd to terminate" on every
/etc/init.d/pcp (re)start. Note that the restart does proceed
unproblematically: the only issue is these scary-looking warnings.
This is mostly a problem if you are running QA
--- a/mgmt/pcp/src/pmcd/rc_pcp 2007-10-09 15:35:59.000000000 +1000
+++ b/mgmt/pcp/src/pmcd/rc_pcp 2007-10-09 15:34:08.816409829 +1000
@@ -461,7 +461,7 @@
# On failure, resort to SIGKILL.
#
$ECHO $PCP_ECHO_N "Waiting for PMCD to terminate ...""$PCP_ECHO_C"
- delay=80 # tenths of a second
+ delay=200 # tenths of a second
for SIG in TERM KILL
do
if [ "x$TOKILL" == "x" ]
Dr.Michael("Kimba")Newton kimbrr@xxxxxxx
|