Monitor elements for viewing locks between members

The IBM® DB2® pureScale® Feature adds a number of new monitor elements that you can use to monitor locking in a DB2 pureScale environment. These monitor elements report specifically on information about lock waits between members in a DB2 pureScale instance.

Locking-related monitor elements

The following monitor elements can be used to get information about lock waits between members and lock escalations:

While not directly related to locking, the following elements can be used to get information about the extent to which things like statements, units of work or workloads in a DB2 pureScale instance wait on a cluster caching facility:

The cf_wait_time element shows how long a statement, unit or work, or workload has had to wait for the cluster caching facility to service a request. Use this element in conjunction with cf_waits (cf_wait_time ÷ cf_waits) to find the average wait time for each request. Generally speaking, the average wait time for requests to be serviced by the cluster caching facility is on the order of a few milliseconds. If you find they are substantially longer (that is, an order of magnitude or more), you might have an InfiniBand setup issue. Alternatively, your cluster caching facility may be overwhelmed with requests that it is not able to keep up with.

Refer to the reference topics for each monitor element to see what monitoring interfaces you can use to examine the data associated with that monitor element.

You can also use the other monitor elements supported by the locking event monitor to view information about locking within members.