The issue actually starts with Alice, as she sends Tx1 with a very low fee. However, Carol (the attacker) exploits this situation further by creating Tx2, a large child transaction with an extremely low fee, which brings down the fee rate of the whole package.

Without Carol's interference, Alice’s Tx1 might get confirmed if someone uses a CPFP method.

In short, Carol’s actions as the attacker are what turn this into a pinning attack.

bitcoin_devs

bitcoin

Transaction Pinning Attack

Is Tx1 actually stuck due to Carol's actions or rather due to the fact that  Alice actually sent Tx1 with a very low fee, considering the known fact that miners will always prefer to go for high fee transactions? In other words, who actually created the problem in the first place? 

I'm curious to learn. Looking forward to your reply on this.  Thank you for this wonderful piece. 