I don't understand how a transaction can write abort statement to log. If system crashes, how will the transaction ever have the time to write abort? In deferred DB modification, you say abort will not be written by transaction (which makes sense). But here in immediate DB modification, you say abort will be written. How?
Thank you very much sir, now it all seems so easy after watching your lectures, you better than the teacher of our college
Dhanyavaad ji.
I don't understand how a transaction can write abort statement to log. If system crashes, how will the transaction ever have the time to write abort?
In deferred DB modification, you say abort will not be written by transaction (which makes sense). But here in immediate DB modification, you say abort will be written. How?
If it checks for commit record in system log , then why to undo again ?????