Welcome to the new Gigaspaces XAP forum. To recover your account, please follow these instructions.

Ask Your Question

Prob killing a replication channel using XAP 7.1 with a static cluster cfg


We have a high availability scenario involving a simple cluster with two spaces - one acting as active, the other as standby. A replication channel is configured from the active to the standby.

When a problem occurs on the active system, the first task in our HA scenario is to stop the replication channel to avoid entry updates related to shutdown activities being replicated from the active to the standby system as the active goes through an orderly shutdown.

We attempt to stop the replication channel by calling IRemoteJSpaceAdmin.changeReplicationState(<target cluster="" member="">, false)

However, We notice that the channel becomes reestablished after 5 seconds. We assume that this is because the repl-find-timeout attribute of the replication cluster configuration is set to 5000 (5 seconds).

Our questions are: Is our assumption correct that the IRemoteJSpaceAdmin.changeReplicationState() method only suspends the replication channel temporarily until the interval specified by repl-find-timeout elapses? If so, is there a better way to permanently kill the replication channel? For example, even though the replication cluster is statically configured, is it possibly to programatically, dynamically remove a node from the cluster.

Thanks for any help,


{quote}This thread was imported from the previous forum. For your reference, the original is [available here|http://forum.openspaces.org/thread.jspa?threadID=3608]{quote}

asked 2011-03-02 05:29:08 -0500

maitiu gravatar image

updated 2013-08-08 09:52:00 -0500

jaissefsfex gravatar image
edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

As far as I remember this might be the case.
I think you should implement the shutdown process in a different manner.
It would be better blocking clients from performing space operations against the active instance once the system going through shutdown process. This will avoid the need to hassle with the replication channel activity that is a delicate component.

To signal a client that the system starting a shutdown process , it should register for notification for a special object. Once a client will be getting such a notification it would stop its activity. You could pause the client or terminate it. Once there are no clients, you can shutdown the space.

answered 2011-03-02 21:47:05 -0500

shay hassidim gravatar image
edit flag offensive delete link more


Thanks Shay,

Agree that the ideal solution is to block clients from performing space operations against the active instance while shutdown in progress. We are working towards this but not there yet,

In the meantime, we are learning/investigating ways to minimize the impact of the current set-up. At: http://www.gigaspaces.com/wiki/displa... , I can see that its possible to override the configuration of GigaSpace. We do this already at start-up using a gs.properties file.

My question is if you know whether this can be done after the space has been started? Specifically, could we change the value of +cluster-config.groups.group.repl-policy.repl-find-timeout+ once the space is already up?

The examples in the documentation seem to indicate that this can only be done before the space is stated

e.g. UrlSpaceConfigurer spaceConfigurer = new UrlSpaceConfigurer("/./space").addProperty("space-config.serialization-type", "1"); IJSpace space = spaceConfigurer.space();

Thanks Matt

maitiu gravatar imagemaitiu ( 2011-03-07 06:25:41 -0500 )edit

You can't change this property value once the space been started. I suggest you approaching GigaSpaces professional services team. They might be able to come up with a solution for you. Shay

shay hassidim gravatar imageshay hassidim ( 2011-03-08 02:36:14 -0500 )edit

Thanks Shay, We are looking at our design options. Just wanted to confirm that last point. You've been very helpful. Matt

maitiu gravatar imagemaitiu ( 2011-03-08 05:43:09 -0500 )edit

Hi Matt

Best way to handle replication is to use replication filters. When space is getting shutdown you can have replication filter to drop all out going replication objects.

For more information http://www.gigaspaces.com/wiki/displa...

Thanks Venkat

venkatg gravatar imagevenkatg ( 2011-03-25 12:53:59 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower


Asked: 2011-03-02 05:29:08 -0500

Seen: 36 times

Last updated: Mar 02 '11