Class Ferret::Store::Lock
In: ext/r_store.c
Parent: Object

A Lock is used to lock a data source so that not more than one output stream can access a data source at one time. It is possible that locks could be disabled. For example a read only index stored on a CDROM would have no need for a lock.

You can use a lock in two ways. Firstly:

  write_lock = @directory.make_lock(LOCK_NAME)
  write_lock.obtain(WRITE_LOCK_TIME_OUT)
    ... # Do your file modifications # ...
  write_lock.release()

Alternatively you could use the while locked method. This ensures that the lock will be released once processing has finished.

  write_lock = @directory.make_lock(LOCK_NAME)
  write_lock.while_locked(WRITE_LOCK_TIME_OUT) do
    ... # Do your file modifications # ...
  end

Methods

Classes and Modules

Class Ferret::Store::Lock::LockError

Public Instance methods

Returns true if the lock has been obtained.

Obtain a lock. Returns true if lock was successfully obtained. Make sure the lock is released using Lock#release. Otherwise you‘ll be left with a stale lock file.

The timeout defaults to 1 second and 5 attempts are made to obtain the lock. If you‘re doing large batch updates on the index with multiple processes you may need to increase the lock timeout but 1 second will be substantial in most cases.

timeout:seconds to wait to obtain lock before timing out and returning false
return:true if lock was successfully obtained. Raises a Lock::LockError otherwise.

Release the lock. This should only be called by the process which obtains the lock.

Run the code in a block while a lock is obtained, automatically releasing the lock when the block returns.

See Lock#obtain for more information on lock timeout.

timeout:seconds to wait to obtain lock before timing out and returning false
return:true if lock was successfully obtained. Raises a Lock::LockError otherwise.

[Validate]