Cluster performance degradation
Hello,
I've been testin cluster for quite some time now, I noticed strange behaviour:
with 2 NDB nodes and 1 API instance (running on one of the NDB nodes) I get 1400 ops/sec when doing SELECT. However with only one node running API and NDB I get 1700 ops/sec with exdactly the same test tool and configuration. Any ideas why this happens.
Same behaviour is noticed when doing INSERT's, just at a slightly bigger scale (2200 vs 2600 ops/sec).
These figures I got on two Dell servers each equiped with two Intel Xeon 2.4GHz cpu and interconnected with Gbit link. Test tool ran on a dedicated server (P3, 977MHz) and had a 100Mbit uplink to the csluster.
I'm going to test 4 node config, but unfortunatelly, I will have only 100Mbit interlink then.
Thanks a lot in advance.
Subject
Views
Written By
Posted
Cluster performance degradation
4059
October 21, 2004 11:01AM
2801
October 23, 2004 01:45PM
Sorry, you can't reply to this topic. It has been closed.
Content reproduced on this site is the property of the respective copyright holders.
It is not reviewed in advance by Oracle and does not necessarily represent the opinion
of Oracle or any other party.