Macro | Brief description |
LKCAN | Cancels lock requests |
LKCVT | Converts lock requests |
LKDEQ | Releases lock requests |
LKENQ | Generates a lock |
LKEQU | Generates DLM-specific layouts |
LKINF | Provides information about locks |
LKLSB | Generates the layout of the Lock Status Block |
The Distributed Lock Manager (DLM) enables tasks running on different nodes of an XCS cluster (XCS = cross-coupled system) to serialize their accesses to shared resources (e.g. files, data records, database blocks, devices, etc.). The resources which are concealed behind a lock are not visible for the DLM: consultation is the responsibility of the users.
The DLM functionality is comparable to the serialization of tasks on a single (local) node (computer). However, it is now also possible to choose between local node locks, i.e. locks which are only visible at the local computer, and cluster locks which are visible at all the computers (= nodes) in the XCS network. To set up an XCS network it is necessary to purchase the product HIPLEX MSCF [26 (Related publications)].
In addition, DLM provides a number of lock modes which permit not only a serialization of (exclusive) access but also, for example, parallel, simultaneous reads by different users.
As far as the DLM is concerned, a lock is identified by
a freely definable lock name assigned by the user
the local scope (SCOPE operand)
the global scope (NAMRNGE operand)
These specifications are used to assign the different calling tasks the lock with which they want to work. Only if all three specifications are the same are two tasks referring to the same lock.
To simplify processing, the calling task receives a lock ID which is then used to identify the lock in all subsequent calls.
DLM locks are not linked to locks of any other lock manager. To ensure that a resource is correctly protected, all users of that resource must use the same lock mechanism.