I don't think there was any disagreement in the fact that 2MB block sizes are good for scaling immediately.

It does introduce some amount of centralization in the sense that it will require more compute/storage to process and maintain the blockchain. But, i feel the tradeoffs at this point were not as significant as it has been shown in media.

However, the block size is best at which the network agrees. If everyone agrees on 2 MB or 4MB, is really just a small battle.
That said, the people who were against the 2MB upgrade had the following argument.

If we do a 2MB upgrade right now, we will point to this upgrade to say 'hey, the last upgrade worked fine, we can double the block size again'.
And if we keep increasing the block size, then the centralization becomes a real issue.

The reason why the 2MB upgrade was not done was to fix this problem in a future proof way - the layer 2 solution.