Payara to AWS cloud issue

92 views
Skip to first unread message

Juha Ranta

unread,
Jun 28, 2018, 2:05:22 PM6/28/18
to Payara Forum
I've been evaluating and testing getting Payara to AWS Cloud with clustering. Currently I have a standard Payara 4 server with a single main application. 

This article seems to detail pretty well what I want to do with Hazelcast AWS Native Discovery plugin: https://blog.payara.fish/aws-native-discovery-with-payara-micro 

I tried the approach with Payara 5 server version, I copied hazelcast.xml, hazelcast-aws.jar, etc like in the demo package to the server directory, etc, but got stuck with class loader issues, class not found exceptions, etc. I also read from some bug report that there were some changes related to serialisation in some 4.* version to prevent this type of issues. Then there's Payara 5 with data grid which claims to help with connecting to things such as AWS cloud with no issues, but it's a new feature and I couldn't find good guidelines on how to do this in this type of configuration. 

If I want to get something like what Hazelcast AWS Native Discovery plugin provides with Payara, in short term should I investigate more Payara 5 data grids or stick with Payara 4 and go along the lines described in that blog post? In either case, should I rather try Payara Micros instead of Servers? I realize they Micros and Services may have different class loader hierarchies, which may help with finding classes.

Ondro Mihályi

unread,
Jun 28, 2018, 2:41:18 PM6/28/18
to Juha Ranta, Payara Forum
Hi,

I would definitely go with Payara Server or Payara Micro 5. It provides the same functionality if you want to use Hazelcast-based cluster but has better support for running in the cloud. Payara 5 datagrids is basically just an improved Payara 4 Hazelcast based clustering.

Payara 5 datagrid clustering is based on DAS (admin server) - Payara provides a custom Hazelcast plugin to connect to the cluster via the Admin server. The advantage is that you don't need to add server instances from the admin server, they can just connect and register to the cluster when you start them. You can start as many instances as you want, either Payara Server instances or Payara Micro, without changing DAS configuration. You still need a DAS server running and the instances have to know its IP address and port. It works well in any cloud environment. The only drawback is that the DAS is not clustered and if it's down you have to start it so that other instances can join the cluster. Docs about Payara Server DAS clustering is here: https://docs.payara.fish/documentation/payara-server/hazelcast/discovery.html
Docs about Payara Micro connecting to DAS clustering is here: https://docs.payara.fish/documentation/payara-micro/clustering/autoclustering.html

If you want autodiscovery without DAS, you need to use the AWS Native Discovery, which you can use also with Payara Server 5 and Payara Micro 5.

All the best
Ondro



--
You received this message because you are subscribed to the Google Groups "Payara Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to payara-forum+unsubscribe@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/payara-forum/01c9967f-710d-4c4d-ba51-caa14290c110%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages