Checkpoints in Data Recovery

Question

Are they in the syllabus? I believe not since nothing is specified in the syllabus but I've seen questions in made easy test series regarding undo and redo list.

$(Start, T_4); (Write,T_4,y,2,3); (Start,T_1);(Commit,T_4);(Write,T_1,z,5,7)\\ (Checkpoint)\\ (Start,T_2);(Write,T_2,x,1,9);(Commit,T_2);(Start,T_3);(Write,T_3,z,7,2);\\ (Crash)$

Give contents of Undo and Redo list.

Comments

Its a part of Recovery system in Transactions regulated by Recovery manager : Redo(Operations of T2 as it commits) and Undo the operations of (T1 and T3) .
PS: T4 has already committed before checkpoint and crash therefore no operations needed.

---

Checkpoint declares a point before which the DBMS was in consistent state, and all the transactions were committed.

So why $T_1 $ needs to be on Undo list. It must have been already committed.

---

Not necessarily ! The log is replicated to disk data i.e updates temp variables value to disk of only those transactions which committed before checkpoints.
Suppose if value of Z updated in disk and after checkpoint T2 fails then z value will become inconsistent therefore manager updates the committed transaction values.

---

Thanks brother.