Bug #64713 NDB Memcached Performance Regression
Submitted: 21 Mar 2012 2:05 Modified: 10 Oct 2013 16:27
Reporter: John Duncan Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S3 (Non-critical)
Version:7.2.2 - 7.2.5 OS:Any
Assigned to: CPU Architecture:Any
Tags: Memcache

[21 Mar 2012 2:05] John Duncan 
Description:
In MySQL Cluster 7.2.2 through 7.2.5, performance of NDB Memcache with a workload that consists heavily of primary key reads is degraded, compared to the first preview release.  

How to repeat:
This can be seen using memslap; for instance, operations per second will decrease from "memslap -c 4" to "memslap -c 6" and level out precisely.
[10 Oct 2013 16:27] Jon Stephens 
Fixed in NDB 7.2+. Documented fix in the NDB 7.2.6 changelog as follows:

      The performance of ndbmemcache with a workload that consisted 
      mostly of primary key reads became degraded.

Closed.