Self Doubt

Question

Is the following schedule conflict serializable

T1	T2	T3
W(X)
commit
	R(X)
	W(X)
		R(X)
		W(X)
		commit
	W(X)
	commit

 

Comments

yes becoz after t3 commit t2's write came

---

does commit matters while checking serializability

@Shaik Masthan

@Mk Utkarsh

---

Manas Mishra

Answer here is wrong?

---

https://gateoverflow.in/112238/self-doubt

check this one also ..answer by bikram sir

---

might be i m wrong , so then is it not serializable ?

---

YES.

It is not a recoverable schedule but it is conflict serializeable schedule.

we don't compare after commit operation. Like R3(X) W3(X) will not be compared to the second W2(X) because T3 is committed and is no more present in the system.

 

---

According to navathe book def of conflict operations definition: it dosn’t say anything about commit of transactions

---

https://gateoverflow.in/1988/gate2014-2-29

---

Satbir what about it?

---

@Shaik Masthan

It will be helpful if you explain

are answers in this links wrong ?

https://gateoverflow.in/81125/conflict-serializability-when-schedule-contains-commit

https://gateoverflow.in/112238/self-doubt

---

@Mk Utkarsh

read the comments in that question. The concept used here is similar to the question mentioned. one of the comment is

Wiki says:->>Another definition for conflict-serializability is that a schedule is conflict-serializable if and only if its precedence graph/serializability graph, when only committed transactions are considered, is acyclic (if the graph is defined to include also uncommitted transactions, then cycles involving uncommitted transactions may occur without conflict serializability violation).

---

Satbir yes that means transactions containing abort are ignored

---

it will not conflict serializable

https://gateoverflow.in/266549/raghuramkrishnan

---

@srestha

the question which you have mentioned is having commit and abort operations at the end. first cycle is forming and then commit and abort are taking place.

but in this question cycle is not forming between T2 and T3 because of the commit operation of T3.

Can you please explain or give reference why we don't see commit operation in case of conflict serializability ?

---

@Arjun sir,

it will be hepful if you clarify this ..lot of confusion :(

---

As per my knowledge, COMMIT means, that transaction completes, then how can it will be conflict with others after it's commit ?

I read that ( From NPTEL lectures ) , there are some irrecoverable schedules which are serializable.

This question is the example for that type !

---

Yes ...correct  @Shaik Masthan

---

Shaik Masthan link of that NPTEL lecture?

---

@Shaik Masthan

where did u read this??

I read that ( From NPTEL lectures ) , there are some irrecoverable schedules which are serializable.

 

---

At 19:00

 

---

T1	T2
R(x)
W(x)
	R(x)
	W(x)
	Commit
Commit

This is irrecoverable but conflict serializable but we are discussing something else na 

---

Bro, If conflict serializable then we can give  the order.. like

T1->T2->T3 but it is conflicting.

T1->T2->T3 and this is also conflicting.

then what is the effect of considering commit in this??

---

if it is View serializable, then exactly we can't give some order, right ?

even in Thomos write Rule, we ignore obsolate writes, right ?

@Mk Utkarsh

That's why i specified as per my knowledge :)

 

Then how this gate question is serializable ? https://gateoverflow.in/1988/gate2014-2-29

---

Because no cycle 

---

Correct me if I am wrong. 

@Shaik Masthan

@Mk Utkarsh

the schedule is not C.S as  we have a cycle in it(T2 , T3).. Commits dont' matter.

only aborts do as we don't count them while makeing dependency graph. 

---

@Shaik Masthan

For that gate 2014 questions ..the schedule is serializable because no cycle in graph ..its so happened coincidently that if we consider commit or not ..in both cases no cycle for that question ...diagram given by @Mk Utkarsh

is correct i think ..

 

Answer 1

Serializable but not recoverable..