Is the following scenario possible with GigaSpace?
(1) Start multiple containers, deploy a PU jar version 1 to all containers.
(2) Prepare PU jar version 2.
(3) Somehow (how?) tell GigaSpace to deploy PU jar version 2 in case if any of already running containers is restarted.
The result will be that the PU version 1 continue to run as long as containers are running, but if a particular container is terminated and restarted, it will pick up the new version of the PU. This will allow to restart containers one-by-one, effectively upgrading the clustered system to a new PU version without a complete restart (of course, assuming that versions are compatible). This would be very useful for a scenario when we want to install an emergency patch, but we do not want to take the whole system down.
↧