i'm trying to figure out how to use/test the lockforupdate correctly, but i found is not function like what i expected
this is just testing
public function index() {
return dd(\DB::transaction(function() {
if (\Auth::guard('user')->check()) {
$model = \App\Models\User::find(1)->lockForUpdate();
sleep(60);
$model->point = 100000;
$model->save();
} else {
$model = \App\Models\User::find(1);
$model->point = 999;
$model->save();
}
return $model;
}));
}
i try to test in 2 browser, browser 1 user logged in and browser 2 not logged in, browser 1 hit refresh, then there will lockforupdate and sleep 60 seconds before update
in the 60 seconds, i go browser 2 and hit refresh, however the record is not locked, i check phpmyadmin and the record is updated(within the 60 seconds lock trigger by browser 1)
but after 60 seconds, the record has been modified again by browser 1(Point 100000)
so am i misunderstanding the lockforupdate is use for?or i test it incorrectly?
what i expected is the row shouldn't be modified by browser 2 in the first 60 seconds(blank page with loading favicon or error throw?)
https://laravel.com/docs/5.2/queries#pessimistic-locking
and i did some research but still cannot understand what different between sharedLock(LOCK IN SHARE MODE) and lockForUpdate(FOR UPDATE)
btw i confirmed the database is innodb
The pessimistic locking is easy to implement using DB transaction as below. In Laravel, all the DB operations between beginTransaction and commit are guaranteed to be committed atomically. Pessimistic Transaction. And, the optimistic locking is, further, implemented as below.
A “for update” lock prevents the selected records from being modified or from being selected with another shared lock. This is what we want. If we run lockForUpdate in our find() statements, they will not be selected by another shared lock.
Pessimistic lockingAs soon as one user starts to update a record, a lock is placed on it. Other users who attempt to update this record are informed that another user has an update in progress. The other users must wait until the first user has finished committing their changes, thereby releasing the record lock.
From Laravel documentation: A shared lock prevents the selected rows from being modified until your transaction is committed. So as it is written - lock will be active from when you call it until your transaction is done.
This work, finally, but still don't understand what sharedLock(LOCK IN SHARE MODE) and lockForUpdate(FOR UPDATE) different
public function index() {
return dd(\DB::transaction(function() {
if (\Auth::guard('user')->check()) {
$model = \App\Models\User::lockForUpdate()->find(1);
sleep(30);
$model->point = 100000;
$model->save();
} else {
$model = \App\Models\User::lockForUpdate()->find(1);
$model->point = $model->point + 1;
$model->save();
}
return $model;
}));
}
So this is old question but I believe my answer can clarify how ->lockForUpdate()
works
From Laravel documentation:
A shared lock prevents the selected rows from being modified until your transaction is committed.
So as it is written - lock will be active from when you call it until your transaction is done.
Remember:
->find(1)
works like ->first()
, ->get()
, ->insert()
, ->save()
etc. - it executes the query
->lockForUpdate()
works like ->where()
, ->select()
, join()
etc. - it adds to the query, but doesn't execute it
$model = \App\Models\User::find(1)->lockForUpdate();
- you try to add lock after query has already executed
$model = \App\Models\User::lockForUpdate()->find(1);
- you add lock before query is executed therefore lock is active until transaction is finished
Difference is that in 1st scenario ->lockForUpdate()
wasn't executed when you taught that it was
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With