This is more of a big data question(GFS related). I'm not getting how checkpointing help? Can you share 1 example(which could be DBMS example).
Say these are the operations on the database.
Time| Operations 0 start T1 1 read A 2 write B=B+20 3 read B 4 write A=A+40 5 Commit T1 6 start T2 7 write C=C+A 8 Write A=A+B 9 Read B 10 Commit T2
I don't get what checkpointing helps in replaying effort here? Can you clarify me?
Checkpoint in dbms can be considered same we create restore point In Windows backup.
Suppose we have checkpoint1 and we want to execute some query and create checkpoint2
then if something goes wrong after creating checkpoint2, we can rollback changes back to the changes and go back to checkpoint1 again.
You can also refer
https://www.csee.umbc.edu/courses/461/current/burt/lectures/lec25/recovery.html (Check log based recovery)
Subscribe to our weekly Newsletter & Keep getting latest article/questions in your inbox weekly