Forum rebuild [almost] complete!

[NEOWatcher]

You may have heard this before, but this is exactly what a test server is for. To be able to go live where things can break and fixes can be rehearsed.

While I completely agree with you, there are some issues that do not appear on a test server.

Unless your test server is a complete clone of the production server, and relatively close to the production data, it's very hard to predict a roll out.

Even then, a rollout runs into unexpected problems. About the only thing that can be done is complete cloning. This would take time and leave us down while the data is moved to the test system, tested and back to production. Half dozen of one, 6 of the other.
Even then, server links and similar items can cause issues.

This is why communication is the key. "we are upgrading, it is a large amount of code", "code is in, but please keep activity to a minimum while we verify the system", etc.

From not knowing the background of this environment, and one who does this for a living, I don't think the update went too horrible except for the communication.

[HenrikOlsen]

As I understand it, the server's in the cloud, which makes cloning it fairly easy and not prohibitively expensive given that the clone only lives for a short time.

[NEOWatcher]

As I understand it, the server's in the cloud, which makes cloning it fairly easy and not prohibitively expensive given that the clone only lives for a short time.

I don't claim to know what they did or how they did it. I was just explaining the problems of a "golive" to show that its hard to expect 100% or close to it.

[Swift]

Extravoice, I moved your post to here