Since the very first version, InnoDB was aborting the whole MySQL or MariaDB server when trying to access a corrupted page.
Earlier, small steps have been taken to improve MariaDB stability, such as
We must enable error handling for corrupted pages instead of allowing the server to crash. This is not only about detecting a checksum failure when reading a corrupted page (like suggested in MySQL Bug#10132 Crashing the server on corrupt InnoDB page is unhelpful, filed on April 25, 2005), but also about performing some operations on corrupted pages whose checksum appears valid.
In some background operations, such as the purge of committed transaction history, errors may be reported to the server error log only, or silently ignored. While executing SQL, errors should be reported back to the client, and the transaction could be rolled back.