Reinstall Effort vs. Restoring a server image related to migrating to the Cloud.

jimmymac

Reputable Poster
Greetings all, we are on E1 9.0, tools 9.1.4.7.

Our company is embarking on a project to move all processing to the Cloud (Amazon AWS). JDE is just one of 50+ applications that we used so we are just one small piece of the puzzle. But in relation to that move, I would appreciate input for the following.

We have six JDE related servers, all running MS 2008 R2. Three of these are used to run weblogic (2 prod, 1 non-prod), one production enterprise server, one development enterprise server and one deployment server.

We have two options to move our processing from the current mode of operation to the cloud. The project team for this migration is preferring that all applications do a reinstall on the new cloud servers, although another option would be to a lift and shift (as they are calling it), basically creating a server image then restoring that image on the cloud server.

My CNC skills are meager, so I'm not sure what is involved in either options, especially the reinstall.

What would the level of effort be for doing a reinstall of these six servers vs. the level of effort for doing to virtual image restore.

We will looking to bring in a senior cnc consultant to assist in some of the planning, as well as serving as a resource as we migrate to the cloud, but initially we are looking for some ideas as to which method (reinstall or lift/shift) would be more cost effective. The project team prefers a reinstall but are open to the other on a case by case basis.

Any input would be appreciated.
 
If you are using VMware vSphere, you can use the AWS Management Portal for vCenter to import your VMs. Very quick and easy. You can also use the AWS CLI to import the images. No need for a reinstall.
 
Hi Jimmy,

At a high level you have three options , though there can be variations of each of these.

1 - export images of your existing servers and import them into AWS . If your machines are physical now , then a P2V step would also be involved. - All server names are kept the same, core JDE would need minimal changes

2 - reinstall servers , but keep the same server names . You spin up brand new machines on AWS . Install the required softwares and then copy / restore all your data


3- reinstall servers , new server names . You spin up brand new machines on AWS . Install the required softwares and then copy / restore all your data. Run scripts to update the server name changes .

In all 3 you will build new full packages etc.


No matter what the approach you ideally want to have both systems up in parallel so you can test extensively and then just refresh your Production Data alone for cutover. When keeping the server names the same this requires some creativity and configuration on the network side (think VLANs etc.) Changing server names on the other hand makes it easy from a network configuration perspective but becomes challenging on all your third party systems and integrations that will talk to JDE especially if those are remaining where they are and not moving with JDE to AWS. Honestly the impact on Third Party integrations and systems is where the challenges usually are. The JDE pieces are fairly easy. Did quite a few of these migrations last year also combined with tools upgrades in some cases.

If you would like to engage professionally please feel free to PM me
 
To add -
In my opinion, you would be better off doing an image restore retaining the same server names or overlaying the current E1 system on a server set with a different name and then updating the server names on the database tables.
 
Back
Top