Hi,

If someone uses 2 slon daemons on a same box with different clusternames
but one of the clusternames is contained in the other it makes the
get_pid function fail.

For example :

cluster 1 = my_cluster_X
cluster 2 = my_cluster_X_2

the get_pid does not work if you try to start the slon for cluster 1 if
the one for cluster 2 is already started.

Below is a patch that fix it :

diff -Nur slony1-1.2.15/tools/altperl/slon-tools.pm
slony1-1.2.15.new/tools/altperl/slon-tools.pm
--- slony1-1.2.15/tools/altperl/slon-tools.pm    2007-03-21
16:11:34.000000000 +0100
+++ slony1-1.2.15.new/tools/altperl/slon-tools.pm    2008-12-16
15:17:45.000000000 +0100
@@ -117,7 +117,7 @@
   my $tpid;
   my ($dbname, $dbport, $dbhost) = ($DBNAME[$nodenum], $PORT[$nodenum],
$HOST[$nodenum]);
   #  print "Searching for PID for $dbname on port $dbport\n";
-  my $command =  ps_args() . "| egrep \"[s]lon .*$CLUSTER_NAME\" |
egrep \"host=$dbhost dbname=$dbname.*port=$dbport\" | sort -n | awk
'{print \$2}'";
+  my $command =  ps_args() . "| egrep \"[s]lon .*$CLUSTER_NAME \" |
egrep \"host=$dbhost dbname=$dbname.*port=$dbport\" | sort -n | awk
'{print \$2}'";
   #print "Command:\n$command\n";
   open(PSOUT, "$command|");
   while ($tpid = <PSOUT>) {

-- 
Cyril SCETBON