Performing UAT on Production Server

By ·
Hello all,

I am new to this area and would like to post the following question:

Currently there exist only development and production servers thus is it advisable for UAT to be performed on a production server?

Before the Go-Live phase, a round of UAT had been performed on the development server.

A day before system Go-Live, the team had performed a last round of simple testing on production server. All production data had been backed up before the last round of testing commenced.

I would like to see for feedbacks / advices if the practice, as mentioned above, is encouraged?

Thanks in advance.

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Answers

Collapse -

Off the top of the head answer

by Tony Hopkinson In reply to Performing UAT on Product ...

would be abolutely not.

UAT of development, depends on how much the environment has an impact, and how close your dev setup is the the production one.

Them developer types have a nasty habit of say putting a software update on so they can use a new feature, without mentioning that it's now a pre-requisite of the fix.

I've seen issues that only appeared on a production server, because Admin forgot to mention something they did on it.

Testing in the development environment, finds errors in the software, in live (or better yet a UAT machine with exactly the same set up as Live) environmental impacts.

If your upgrade and backout plan is reliable and it's not too different fom dev, I suppose you could give it a go, but I've seen some 'trivial' difefrences cause major pratfalls on occasion.

I'd want a box that I could set up to be as close to the production server as practical.

Last time I set this up I was on paired servers one live and one backup, so I used to UAT a new version on the backup, If all went well swap it in. Give it a bit to make I sure hadn't missed anything, then upgrade the other.

That was 24/7 manufacturing, I justified the back up machine, with a half hour stoppage, never mind hardware failures and other potential reasons to have a standby machine.

Collapse -

Agree 100%

by robo_dev In reply to Off the top of the head a ...

the only exception would be if a virtual machine (e.g. VMware environment) existed. But then, effectively, it is not the 'same box'.

Otherwise, you never know what can happen.

I once saw a software team basically knock a mainframe offline (over 20K users), because they installed test software on a 'test partition'. Unfortunately, the test software (Linux OS) installed a router component which redirected all external traffic to itself, making the main application unavailable.

Related Discussions

Related Forums