Re: Sandesha2 question?

--0015175d077eeb82ec048c2fbfd5
Content-Type: text/plain; charset=ISO-8859-1

On Fri, Jul 23, 2010 at 4:47 PM, David Parsons1 <[email protected]> wrote:

>
> Hi,
>
> The lastMessage property isn't saying to the Sandesha Handler wait for the
> last message before delivering any of the messages on that sequence to the
> client or service.  All it is saying to Sandesha is that there will be no
> more messages sent on this sequence so you can terminate it.  It is just
> there for sequence administration.  One thing you can do is set the inorder
> property.  This tells the Sandesha handler to deliver the messages that have
> arrived in the order they were originally sent by the client.  Therefore if
> message 1 & 3 have arrived but not message 2, Sandesha would have the
> service invoked by message 1 but not message 3.  It would hold onto message
> 3 until msg 2 had arrived and been invoked.
>
> If you want all or nothing you either need to trust sandesha i.e. that all
> the messages will get sent.  One drawback of this is that your have to rely
> on the server hosting your service to always be available, that way the
> resending of lost messages is handled for you.  Or you switch to having some
> transactional solution.
>
> I have an article that has some early sections that may be of interest.
>  The sections that may be of interest talk about what WS-RM is and the key
> concepts that it uses,
> http://www.ibm.com/developerworks/websphere/library/techarticles/0912_parsons/0912_parsons.html
>

In this article there are three ways of using WS-RM, namely
UnManagedNonPersistent, ManagedNonPersistent and ManagedPersistent.

What is the real benefit of using UnManagedNonPersistent. In your article
there are two things. Here first thing does not apply since it can not
survive on a server restart. Can we achieve the second benefit just
increasing the tcp time out time? In other words can we use the tcp
retransmission? I think WS-RM follows the same principles as in tcp
protocol.

For ManagedNonPersistent and ManagedPersistent what does the clustering
realy means? I assume underling WS-RM implementation that WAS use is
Sandesha2. So it is about sharing one storage between different RM Nodes so
that they can use to receive messages for a same WS-RM sequence? Do you
refer this common storage as messging engine?

I did my Msc[1] on WS-RM to reasearch on how to achive application to
application reliability. Is there a similar thing happens in the
ManagedPersistent using distributed transactions? In my work, it can support
stanalone clients since I uses the Atomikos library get distributed
transactions.

thanks,
Amila.

[1]
http://amilachinthaka.blogspot.com/2010/03/queued-transaction-processing-with-ws.html


>
> Regards,
>
> Dave
>
> Dave Parsons
> Web Services Development
> INTERNAL:  David Parsons1/UK/IBM@IBMGB :: DE3F20 :: 246930
> EXTERNAL:  [email protected] :: (01962) 816930
> Mail Point 211, IBM Hursley Park, Winchester. SO21 2JN
>
>
>
>  From: venkov <[email protected]> To: [email protected]
> Date: 23/07/2010 10:20 Subject: Sandesha2 question?
> ------------------------------
>
>
>
>
> How does Exactly-once option of WSRM works?
>
> I am using the ping sample from sandesha2
> where it sends 3 messages and the 3rd is set to be last!
> after it gets the 3rd one it terminates the sequence and everything is ok!
>
> But if i remove the last message property or
> before i send the last one i terminate the program!
>
> I have the following situation the server gets first and second message,
> execute the action for each one, without waiting for the last message!
>
> I want to execute all or nothing!
> --
> View this message in context:
> http://old.nabble.com/Sandesha2-question--tp29245485p29245485.html
> Sent from the Apache Sandesha mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>
>
>
>
>
>  ------------------------------
>
> *
> *
>
> *Unless stated otherwise above:
> IBM United Kingdom Limited - Registered in England and Wales with number
> 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
> *
>
>
>
>
>
>
>


-- 
Amila Suriarachchi
WSO2 Inc.
blog: http://amilachinthaka.blogspot.com/

--0015175d077eeb82ec048c2fbfd5
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<br><br><div class=3D"gmail_quote">On Fri, Jul 23, 2010 at 4:47 PM, David P=
arsons1 <span dir=3D"ltr">&lt;<a href=3D"mailto:[email protected]">PARSON=
[email protected]</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" st=
yle=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex=
; padding-left: 1ex;">

<br><font size=3D"2" face=3D"sans-serif">Hi,</font>
<br>
<br><font size=3D"2" face=3D"sans-serif">The lastMessage property isn&#39;t=
 saying
to the Sandesha Handler wait for the last message before delivering any
of the messages on that sequence to the client or service. =A0All it
is saying to Sandesha is that there will be no more messages sent on this
sequence so you can terminate it. =A0It is just there for sequence administ=
ration.
=A0One thing you can do is set the inorder property. =A0This tells
the Sandesha handler to deliver the messages that have arrived in the order
they were originally sent by the client. =A0Therefore if message 1 &amp;
3 have arrived but not message 2, Sandesha would have the service invoked
by message 1 but not message 3. =A0It would hold onto message 3 until
msg 2 had arrived and been invoked.</font>
<br>
<br><font size=3D"2" face=3D"sans-serif">If you want all or nothing you eit=
her
need to trust sandesha i.e. that all the messages will get sent. =A0One
drawback of this is that your have to rely on the server hosting your servi=
ce
to always be available, that way the resending of lost messages is handled
for you. =A0Or you switch to having some transactional solution.</font>
<br>
<br><font size=3D"2" face=3D"sans-serif">I have an article that has some ea=
rly
sections that may be of interest. =A0The sections that may be of interest
talk about what WS-RM is and the key concepts that it uses, </font><a href=
=3D"http://www.ibm.com/developerworks/websphere/library/techarticles/0912_p=
arsons/0912_parsons.html" target=3D"_blank"><font size=3D"2" face=3D"sans-s=
erif">http://www.ibm.com/developerworks/websphere/library/techarticles/0912=
_parsons/0912_parsons.html</font></a>
<br></blockquote><div><br>In this article there are three ways of using WS-=
RM, namely UnManagedNonPersistent, ManagedNonPersistent and ManagedPersiste=
nt.<br><br>What is the real benefit of using UnManagedNonPersistent. In you=
r article there are two things. Here first thing does not apply since it ca=
n not survive on a server restart. Can we achieve the second benefit just i=
ncreasing the tcp time out time? In other words can we use the tcp retransm=
ission? I think WS-RM follows the same principles as in tcp protocol.<br>
<br>For ManagedNonPersistent and ManagedPersistent what does the clustering=
 realy means? I assume underling WS-RM implementation that WAS use is Sande=
sha2. So it is about sharing one storage between different RM Nodes so that=
 they can use to receive messages for a same WS-RM sequence? Do you refer t=
his common storage as messging engine?<br>
<br>I did my Msc[1] on WS-RM to reasearch on how to achive application to a=
pplication reliability. Is there a similar thing happens in the ManagedPers=
istent using distributed transactions? In my work, it can support stanalone=
 clients since I uses the Atomikos library get distributed transactions.<br=
>
<br>thanks,<br>Amila.<br><br>[1] <a href=3D"http://amilachinthaka.blogspot.=
com/2010/03/queued-transaction-processing-with-ws.html">http://amilachintha=
ka.blogspot.com/2010/03/queued-transaction-processing-with-ws.html</a><br>
=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px solid =
rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br><font size=3D"2" face=3D"sans-serif">Regards,</font>
<br>
<br><font size=3D"2" face=3D"sans-serif">Dave</font>
<br>
<br><font size=3D"2" face=3D"sans-serif">Dave Parsons<br>
Web Services Development<br>
INTERNAL: =A0David Parsons1/UK/IBM@IBMGB :: DE3F20 :: 246930<br>
EXTERNAL: =A0<a href=3D"mailto:[email protected]" target=3D"_blank">Parso=
[email protected]</a> :: (01962) 816930<br>
Mail Point 211, IBM Hursley Park, Winchester. SO21 2JN<br>
</font>
<br>
<br>
<br>
<table width=3D"100%">
<tbody><tr valign=3D"top">
<td><font size=3D"1" color=3D"#5f5f5f" face=3D"sans-serif">From:</font>
</td><td><font size=3D"1" face=3D"sans-serif">venkov &lt;<a href=3D"mailto:=
[email protected]" target=3D"_blank">[email protected]</a>&=
gt;</font>
</td></tr><tr valign=3D"top">
<td><font size=3D"1" color=3D"#5f5f5f" face=3D"sans-serif">To:</font>
</td><td><font size=3D"1" face=3D"sans-serif"><a href=3D"mailto:sandesha-de=
[email protected]" target=3D"_blank">[email protected]</a></font>
</td></tr><tr valign=3D"top">
<td><font size=3D"1" color=3D"#5f5f5f" face=3D"sans-serif">Date:</font>
</td><td><font size=3D"1" face=3D"sans-serif">23/07/2010 10:20</font>
</td></tr><tr valign=3D"top">
<td><font size=3D"1" color=3D"#5f5f5f" face=3D"sans-serif">Subject:</font>
</td><td><font size=3D"1" face=3D"sans-serif">Sandesha2 question?</font></t=
d></tr></tbody></table>
<br>
<hr noshade><div><div></div><div class=3D"h5">
<br>
<br>
<br><tt><font size=3D"2"><br>
How does Exactly-once option of WSRM works?<br>
<br>
I am using the ping sample from sandesha2<br>
where it sends 3 messages and the 3rd is set to be last!<br>
after it gets the 3rd one it terminates the sequence and everything is
ok!<br>
<br>
But if i remove the last message property or <br>
before i send the last one i terminate the program!<br>
<br>
I have the following situation the server gets first and second message,<br=
>
execute the action for each one, without waiting for the last message!<br>
<br>
I want to execute all or nothing!<br>
-- <br>
View this message in context: </font></tt><a href=3D"http://old.nabble.com/=
Sandesha2-question--tp29245485p29245485.html" target=3D"_blank"><tt><font s=
ize=3D"2">http://old.nabble.com/Sandesha2-question--tp29245485p29245485.htm=
l</font></tt></a><tt><font size=3D"2"><br>

Sent from the Apache Sandesha mailing list archive at Nabble.com.<br>
<br>
<br>
---------------------------------------------------------------------<br>
To unsubscribe, e-mail: <a href=3D"mailto:[email protected]=
e.org" target=3D"_blank">[email protected]</a><br>
For additional commands, e-mail: <a href=3D"mailto:sandesha-dev-help-J5gJY/cNB/[email protected]=
che.org" target=3D"_blank">[email protected]</a><br>
<br>
</font></tt>
<br>
<br><font size=3D"2" face=3D"sans-serif"><br>
</font>
<br><font size=3D"2" face=3D"sans-serif"><br>
</font>
</div></div><hr><font size=3D"2" face=3D"sans-serif"><br>
<i><br>
</i></font>
<p><font size=3D"2" face=3D"sans-serif"><i>Unless stated otherwise above:<b=
r>
IBM United Kingdom Limited - Registered in England and Wales with number
741598. <br>
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU</i></font>
</p><p><font size=3D"2" face=3D"sans-serif"><br>
<br>
</font>
<br>
<br><font size=3D"2" face=3D"sans-serif"><br>
</font></p></blockquote></div><br><br clear=3D"all"><br>-- <br>Amila Suriar=
achchi<br>WSO2 Inc.<br>blog: <a href=3D"http://amilachinthaka.blogspot.com/=
">http://amilachinthaka.blogspot.com/</a><br>

--0015175d077eeb82ec048c2fbfd5--