Re: [rules-users] Workbench and GIT Repo High Availability
Hello Zahid,
Comments inline...
On May 13, 2014, at 3:56 AM, Zahid Ahmed <zahid.ahmed(a)emirates.com> wrote:
Hi,
We are in a process of setting up a centralized drools6 Workbench in our organization.
This way we can keep our assets and artefacts in a central location. We will make this
central Workbench clustered for high availability. I have a few queries in regard to
clustering the Workbench.
Problem Statement
As workbench is using GIT, this means we need to have an enterprise level GIT repository
where all the assets get stored for Workbench. As GIT is file-system based , it in turn,
calls for a clustered GIT server within the organization. Our organization will prefer to
have a private GIT repository hosted within office network which is highly available and
wants to avoid any cost for GIT repository setup.
Workbench doesn't force you to have any extra cost with GIT, as it comes with it and
you don't need an external (outside your network) provider as Workbench has it's
own GIT infra.
1. I want to know how to make Workbench highly available , with GIT also Highly
available. Is chapter# 18 VFS CLUSTERING is the answer to high available GIT and WORKBENCH
both ? or is there any other way to make GIT repository highly available.
The Workbench cluster is a Highly Available type of cluster, all you need is follow the
instructions and you'll have what you need. To be even more clear: as GIT is embedded
in Workbench solution, once you setup a Workbench cluster you get for free a GIT HA
setup.
But if you're looking for a proxy based solution (Workbench accessing your own
external GIT cluster using whatever solution you have) - this we don't support,
Workbench will always have it's own GIT infra embedded.
2. Will ZooKeeper and Helix be enough to make the clustered
environment for both Workbench with its GIT repo ?
Yes... ZK and Helix does the job! Be aware that you'd need to have a proper ZK setup
(another cluster) in order to obtain a real HA cluster. Your final solution will be
probably end up with 3 ZKs in a cluster + the Workbench cluster (both clusters can be
setup in the same nodes if you want).
Regards,
---
Alexandre Porcelli
Principal Software Engineer
Red Hat Business Systems and Intelligence Group
Regards,
Zahid Ahmed
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users