Hi Carolina,
I will ask our system administrators if they have further suggestions, but will start by listing some of the more obvious / basic things to be aware of:
First, it will be important to make sure that each AtoM instance will have enough system resources. As noted in our
Technical Requirements documentation, we recommend the following as minimum targets for a production system:
- Processor: 2 vCPUs @ 2.3GHz
- Memory: 7GB
- Disk space (processing): 50GB at a minimum for AtoM’s core stack plus more storage would be required for supporting any substantial number of digital objects.
I think that you will likely be best served by installing each AtoM instance in its own VM, so that resource management can be scaled as needed, and so there are not conflicts in file paths, dependencies, etc.
That is my second reminder: make sure that you are using unique names, IP addresses, and paths when configuring various elements of each AtoM instance. You might want to call them atom-1 and atom-2, or atom-dev and atom-prod, or similar - and then consistenly use these names throughout the installation process (for example, for the MySQL database name, the Nginx configuration file name, the Elastisearch index name, etc.). You'll need to make sure that IP addresses and ports used don't conflict as well.
In general, if you are avoiding conflicts and providing each instance with enough resources - and in particular, if you are using VMs to manage each instance - then having more than 1 AtoM instance on a single server shouldn't inherently be a problem.
I will see if our Ops team has any further suggestions or warnings to pass on.
Cheers,