Performance Down During LOAD CACHE (Doc ID 1610645.1)

Last updated on MARCH 02, 2016

Applies to:

Oracle TimesTen In-Memory Database - Version 11.2.2.2.0 and later
Information in this document applies to any platform.

Goal

A customer has seen contention reported on index latches while parallel load cache group operations are executing. Index latch waits are seen both on cache tables being PARALLEL loaded and normal tables which should not have any relationship with the cache tables being loaded. Here is an example of errors seen;

19796/10026f170: ConnId=6 (radiusd) waiting for latch "Heap[31] - Index"(108096), Holder=11 (Cache Load) PID 27659, now 1 secs
19796/10026f170: ConnId=6 (radiusd) waiting for latch "Heap[31] - Index"(108096), Holder=10 (Cache Load) PID 27659, now 1 secs
19796/10026f170: ConnId=6 (radiusd) waiting for latch "Heap[31] - Index"(108096), Holder=14 (Cache Load) PID 27659, now 1 secs
19796/10026f170: ConnId=6 (radiusd) waiting for latch "Heap[31] - Index"(108096), Holder=17 (Cache Load) PID 27659, now 1 secs
12858/10026f170: ConnId=60 (radiusd) waiting for latch "Index MY.MADEUPTABLE"(2228400), Holder=6 (myholder) PID 19796, now 1 secs
19796/10026f170: ConnId=6 (radiusd) waiting for latch "Heap[31] - Index"(108096), Holder=9 (Cache Load) PID 27659, now 1 secs
12790/10026f170: ConnId=50 (radiusd) waiting for latch "Index MY.MADEUPTABLE"(2228400), Holder=6 (myholder) PID 19796, now 1 secs

MY.MADEUPTABLE is a normal table having DML run against it. This should not be being touched by the LOAD CACHE GROUPs which occur against different, cache group objects.
 

Solution

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms