Menu
Community /Pin to ProfileBookmark
@NogDogAug 25.2006 — #I would think the easiest solution would be to do the incrementing as part of the query, something like:
@NogDogAug 25.2006 — #My understanding is that MySQL waits until it gets its lock. I'm not sure if MySQL has a timeout limit on that, but if your PHP script times out then the lock request will be dropped as soon as your script dies, since such death will release the database connection. @NogDogAug 28.2006 — #My understanding is that MySQL will wait until either it gets a lock or the process requesting it dies. Basically, each read/write/lock request gets put on a queue, with MySQL deciding [i]where[/i] on the queue it goes depending on what type of request and what type of lock (if any) is currently in place. So your PHP script is ignorant of the MySQL locking: all it sees is different response times from the mysql_query() calls. If for some reason you have a deadlock condition, then eventually your script might time out and die.
Race Conditions when accessing MySQL?
In a new web site I’m building, I may have to create a page that queries the MySQL database for a value in a particular record, increments that value, then queries the database again to update the record with the new value.
My question is: Is there a danger of a race condition here if this page is loaded by two users simultaneously? (i.e., user A’s first query reads the value, then user B’s first query reads it, they both increment, and the final value represents only a single increment after both insert queries) Or do PHP and MySQL already have built-in measures to prevent this?
If it is a danger, what’s a possible solution? And how high does the traffic on my site need to be before I should even bother with one?
Sign in
to post a comment6 Comments(s) ↴
0
@NogDogAug 25.2006 — #I would think the easiest solution would be to do the incrementing as part of the query, something like:[code=php]
$query = <<<EOD
UPDATE table_name SET column_name = column_name + $value WHERE key_column = $id";
[/code]0
@NogDogAug 25.2006 — #0
@NogDogAug 28.2006 — #My understanding is that MySQL will wait until either it gets a lock or the process requesting it dies. Basically, each read/write/lock request gets put on a queue, with MySQL deciding [i]where[/i] on the queue it goes depending on what type of request and what type of lock (if any) is currently in place. So your PHP script is ignorant of the MySQL locking: all it sees is different response times from the mysql_query() calls. If for some reason you have a deadlock condition, then eventually your script might time out and die.