Yesterday we had an EA crash and discovered that one of our most connected objects in the enterprise-wide repository had been trashed! The crash also took out about 20 related connector proxies (yes, , they're connector proxies - proxies for connectors - not proxy connectors - connectors for proxies) and other associated connectors, leaving the repository in a non-integral state.
Fortunately, because of our snapshot process, we have the information to recover and we're currently working on a process to automate this. The reason?
Well, while this instance was caused by Sparx EA crashing, it could just as easily been caused by inadvertent user action!
Consequently, if a request to purge an item from the repository (such as: [Ctrl+Del]) is received by EA and the item is connected above one or more thresholds (such as # of connections, # of diagrams it appears on, etc.) there should be an additional check before EA finally purges the item. For example, you might have to enter the name of the item into a special dialog.
Thoughts?
Reported,
Paolo