Update of /home/cvsd/slony1/slony1-engine/doc/adminguide
In directory main.slony.info:/tmp/cvs-serv11243

Modified Files:
      Tag: REL_2_0_STABLE
	slonik_ref.sgml 
Log Message:
Elaborate on the transactionality that try{} blocks provide...  They do NOT
force downstream processing of events to nest the processing into a single
transaction.

As pointed out by Dmitry Koterov on list


Index: slonik_ref.sgml
===================================================================
RCS file: /home/cvsd/slony1/slony1-engine/doc/adminguide/slonik_ref.sgml,v
retrieving revision 1.92.2.2
retrieving revision 1.92.2.3
diff -C2 -d -r1.92.2.2 -r1.92.2.3
*** slonik_ref.sgml	19 Feb 2009 18:39:50 -0000	1.92.2.2
--- slonik_ref.sgml	2 Apr 2009 18:52:25 -0000	1.92.2.3
***************
*** 84,87 ****
--- 84,104 ----
       <para> Those commands are grouped together into one transaction
        per participating node. </para>
+ 
+      <para> Note that this does not enforce grouping of the actions as
+      a single transaction on all nodes.  For instance, consider the
+      following slonik code:</para>
+      <programlisting>
+      try {
+          execute script (set id = 1, filename = '/tmp/script1.sql', event node=1);
+          execute script (set id = 1, filename = '/tmp/script2.sql', event node=1);
+      }
+      </programlisting>
+ 
+      <para> This <emphasis>would</emphasis> be processed within a
+      single BEGIN/COMMIT on node 1.  However, the requests are
+      separated into two <command>DDL_SCRIPT</command> events so that
+      each will be run individually, in separate transactions, on other
+      nodes in the cluster. </para>
+ 
  <!-- ************************************************************ --></sect3></sect2></sect1></article>