Question

Would you please look at this question and post the detailed and reasoned answers please? The sooner is better. Thanks

5.2) (5 Points) Following is an example of banking transaction. In each case of (a), (b), (c) transaction to hits a failure. Based on the step in each scenario, an Undo or Redo, or a combination of recovery steps are performed. <To start > <TO, A, 800, 450 > <To, B, 765, 980 > <To start > <TO, A, 800, 450 > <To, B, 765, 980 > <To commit > <T 1 start > <T1, C, 1700, 1600 > <T 1 commit> (a) <To start > <TO, A, 800, 450 > <TO, B, 765, 980 > <To commit> <T 1 start > <T, C 1700, 1600 > (b) What will be the value of? List what recovery actions were performed for each step (Undo/Redo). (a) A, B and C (b) A, B and (c) A and B

Answer 1

Ans:

The following rules would needed to be followed so that the lost transaction may be recovered either through redo or undo actions which may be alternatively or combinely used to revert the failure and transfer the system to a consistent state once more. For the above process following steps should be followed:

Undo - when a log record Ti is undoed the log record <Ti, X, V1, V2> writes the old value of V1 to X. undo(Ti ) restores the value of all data items updated by Ti to their old values, going backwards from the last log record for Ti. and each time the above undo operation takes place it is mentioned (added) to the log records denoted by transaction:

<Ti, X, V>. and is followed by <Ti abort> as soon as the transaction is complete.

Redo - while  Redo of a log record writes the new value V2 to X. redo(Ti ) sets the value of all data items updated by Ti to the new values, going forward from the first log record for Ti. But is not followed by any logging.

When to use which?

When recovering after a failure:

– Transaction Ti needs to be undone if the log

• contains the record <Ti start>,

• but does not contain either the record <Ti commit> or <Ti abort> .

– While Transaction Ti needs to be redone if the log

• contains the records <Ti start>,

• and contains the record or <Ti commit> or <Ti abort> .

So, on using the above operations we have:

a). Operations to be followed :redo (T0 ) and redo (T1 ): (Since both T0 and T1 were already commited)

Since the redo phase is not added in the log therefore the log in scenario a) would just be redoed:

<T0 start>
<T0, A, 800, 450>
<T0, B, 765, 980>
<T0, commit>
<T1 start>
<T1, C, 1700, 1600>
<T1 commit>

Values of : A and B are set to 450 and 980 respectively. Then C is set to 1600.

b). Operations which would be followed : redo (T0 ) and undo (T1 ):

Log records <T1, C, 1700>, <T1 abort> are written out.

<T0 start>
<T0, A, 800, 450>
<T0, B, 765, 980>
<T0, commit>
<T1 start>
<T1, C, 1700, 1600>
<T1, C, 1700>
<T1 abort>

Values of : A and B are set to 4500 and 980, and C is restored to 1700.

c). Operations which would be followed : undo (T0 ):

and log records <T0, B, 765>, <T0, A, 800> and <T0 abort> are written out

<T0 start>
<T0, A, 800, 450>
<T0, B, 765, 980>
<T0, B, 765>
<T0, A, 800>
<T0 abort>

Values of : B is restored to 765 and A to 800.

( PLEASE DO LIKE AND UPVOTE IF THIS WAS HELPFUL!)

( THANK YOU SO MUCH IN ADVANCE!)