A question about Optimistic Locking

Hello all,for Optimistic Locking, I believe
the implementation is like this:
rails will issue the following statement to database
Update user set name=“a”,version=2 where id=1 and version=1
but the problem is, generally database will have transaction Isolation
(I believe Oracle do),
so when Thread 1 issue that statement to database but not commited,
Thread 2 will also issue that statement to database, because transaction
Isolation,
thread 2 will not see this user as version 2 but still version 1(unless
thread 1 commited)
(otherwise thread 2 is reading uncommitted which is not correct),
then thread 1 commit, thread 2 commit,
then all of a sudden, all threads are happy, the user record is still
being
modified by
two threads, the optimistic locking doesn’t work,

Is this correct?
Correct me if I’m wrong.

http://hi.baidu.com/femto
https://github.com/femto

On Monday, March 3, 2014 3:37:06 AM UTC, femto wrote:

thread 2 will not see this user as version 2 but still version 1(unless
thread 1 commited)
(otherwise thread 2 is reading uncommitted which is not correct),
then thread 1 commit, thread 2 commit,
then all of a sudden, all threads are happy, the user record is still
being modified by
two threads, the optimistic locking doesn’t work,

Is this correct?
Correct me if I’m wrong.

Don’t know about oracle but for mysql to update the row you’d have to
acquire a lock on it (an exclusive lock if my memory is correct) - the
second thread would have to wait for the first transaction to be
committed
in order to be able to make its update

Fred

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs