Examine the following details for a six-Instance RAC database:
What is the most prominent problem indicated by the above output?
A.
high input/output (I/O) delays
B.
CPU saturation and memory depletion
C.
large number of untuned queries from one of the nodes
D.
misconfigured or faulty interconnect
Explanation:
In Oracle RAC environments, RDBMS gathers global cache work load statistics
which are reported in STATSPACK, AWRs and GRID CONTROL. Global cache lost blocks
statistics (“gc cr block lost” and/or “gc current block lost”) for each node in the cluster as well as
aggregate statistics for the cluster represent a problem or inefficiencies in packet processing for
the interconnect traffic. These statistics should be monitored and evaluated regularly to guarantee
efficient interconnect Global Cache and Enqueue Service (GCS/GES) and cluster processing. Any
block loss indicates a problem in network packet processing and should be investigated.
Symptoms:
Primary:
Secondary:
Changes
As explained above, Lost blocks are generally caused by unreliable Private network. This can be
caused by a bad patch or faulty network configuration or hardware issue.
Cause
In most cases, gc block lost has been attributed to (a) A missing OS patch (b) Bad network card
(c) Bad cable (d) Bad switch (e) One of the network settings.
Oracle Metalinkgc block lost diagnostics [ID 563566.1]
A is wrong. high I/O won’t cause gc wait
B is wrong.
C is wrong.
D is right. gc cr block busy/lost are interconnect related.
So D. is correct