[1.2.1 → 1.2.8] Is copying full FE/BE directories a viable backup method for offline upgrade?

[h4work]

I'm currently using Apache Doris version 1.2.1. Due to encountering a bug caused by DELETE statements, I plan to upgrade to version 1.2.8.

Since I don't have access to remote storage for data backup, my plan is as follows:

1. Perform a complete system shutdown of the Doris cluster.
2. Copy the entire FE (Frontend) and BE (Backend) directories to a local disk as a backup.
3. Proceed with the upgrade to 1.2.8.
4. If any issues arise during the upgrade, roll back by starting the cluster directly using the backed-up FE/BE directories.

Hardware Configuration: Each node has 2TB of storage space, with approximately 400GB used and ample free space remaining.

I'm unsure if this backup/rollback method is feasible. Has anyone in the community tried this approach before? Any advice or feedback would be greatly appreciated!