-
1. Re: confirmation-window-size in hornetq-jms.xml
pinkushn Feb 20, 2010 10:30 PM (in response to pinkushn)Hmmm... Now I see in the source code that the default is indeed -1. Is the manual correct in stating that that "Setting this parameter to -1 disables any buffering and prevents any re-attachment from occurring?" If so, it seems odd that the Reattach example does not change the default of -1.
-
2. Re: confirmation-window-size in hornetq-jms.xml
timfox Feb 21, 2010 5:30 AM (in response to pinkushn)Whether you need to set this or not depends on what you're trying to achieve, and I'm not sure what that is.....
-
3. Re: confirmation-window-size in hornetq-jms.xml
pinkushn Feb 21, 2010 2:21 PM (in response to timfox)I am hoping to have remote clients quietly reattach when networking breaks temporarily. I would like any client commands that were not transacted on the server prior to the break to be transacted after the reattachment. I think my hornetq-jms.xml should read something like this:
<retry-interval>4000</retry-interval> <reconnect-attempts>10</reconnect-attempts> <confirmation-window-size>1000000</confirmation-window-size>
Is that correct?
-
4. Re: confirmation-window-size in hornetq-jms.xml
jbmuser Mar 29, 2010 8:49 AM (in response to pinkushn)I am facing this exact same problem. Statement in User Manual is confusing. So is the Example.
All I want is my clients to re-attach (without triggering ExceptionListeners) to server on transient network failures (e.g. some delay in response, ping pong/packet loss etc ). What are the changes I need to make in JMS configurations. Appreciate your help on this.
-
5. Re: confirmation-window-size in hornetq-jms.xml
jbmuser Apr 7, 2010 6:44 AM (in response to jbmuser)Any advice on this?
-
6. Re: confirmation-window-size in hornetq-jms.xml
timfox Apr 7, 2010 6:57 AM (in response to jbmuser)Bijith Kumar wrote:
Any advice on this?
What Josh Britton said is correct.
-
7. Re: confirmation-window-size in hornetq-jms.xml
timfox Apr 7, 2010 6:59 AM (in response to timfox)The default value is -1, i.e. no automatic reconnection.
The statement in the user manual that it's 1 MiB is a typo, this has already been discussed on other threads.
-
8. Re: confirmation-window-size in hornetq-jms.xml
jbmuser Apr 7, 2010 7:37 AM (in response to timfox)Thank you Tim for the clarification. I have another question though. If the deafult value for confirmation-window-size is -1 (means re-attachment disabled), then how does the reattach-node example work with this default value. I don't seee this example change deafult value anywhere (neither using API nor in hornetq-jms.xml)
-
9. Re: confirmation-window-size in hornetq-jms.xml
timfox Apr 7, 2010 8:01 AM (in response to jbmuser)If it can't reattach it will try reconnecting.
For the difference between reconnection and reattachment see this chapter:
http://hornetq.sourceforge.net/docs/hornetq-2.0.0.GA/user-manual/en/html/client-reconnection.html
-
10. Re: confirmation-window-size in hornetq-jms.xml
jbmuser Apr 7, 2010 9:04 AM (in response to timfox)I understand teh difference. However, the purpose of re-attach example is to demonstrate re-attach and not reconnect. Below statement is from the description of re-attach example
When the client reattaches to the server it will be able to resume using its sessions and connections where it left off
This is different to client reconnect as the sessions, consumers etc still exist on the server. With reconnect The client recreates its sessions and consumers as needed.
So it seems the example is wrong. i.e. it should have changed the confirmation-window-size from default -1. This caused the confusion.
-
11. Re: confirmation-window-size in hornetq-jms.xml
timfox Apr 9, 2010 6:09 AM (in response to jbmuser)Looks like the example wasn't updated when the confirmation-window-size default was changed from 1 MiB to -1