Oracle 5.6 Vs Percona 5.6: High CPU Utilization in Percona
Though it is not directly related to Oracle Mysql 5.6, I am guessing some one can help me moving my inestigation in right direction:
We are in process of upgrading our (Percona) Mysql 5.5 mysql servers to Mysql 5.6. We have upgraded all our slaves except one.
However, we have observed that for the same workload, 5.6 slaves are consuming more user CPU.
CPU utlization on 5.5 slave continues to be lower. For comparison, we have setup an additional slave on Oracle's community Mysql 5.6. Here the cpu utilization is lower.
Does the percona Mysql 5.6 consumes more cpu than previous version (5.5) or Oracle community 5.6 ? Has anyone else observed similar behavior? In our environment this observation is consistent and reproducible. Certainly moving to Oracle Mysql 5.6 is one solution, let me know if anyone has idea how to resolve this or how to investigate it further. (Tracing mysqld daemon has not given any clue.)
Slave upgraded to Percona 5.6 (EST Timezone)
01:20:01 PM CPU %user %nice %system %iowait %steal %idle
01:20:01 PM all 10.97 0.00 0.05 0.19 0.00 88.79
01:30:01 PM all 11.20 0.00 0.06 0.01 0.00 88.73
01:40:01 PM all 9.48 0.00 0.08 0.02 0.00 90.41
Slave still on 5.5:
10:20:01 AM CPU %user %nice %system %iowait %steal %idle
10:20:01 AM all 0.26 0.00 0.04 0.39 0.00 99.31
10:30:01 AM all 0.23 0.00 0.10 0.08 0.00 99.58
10:40:01 AM all 0.32 0.00 0.05 0.17 0.00 99.47
Slave upgraded to Oracle Mysql 5.6:
10:20:01 AM CPU %user %nice %system %iowait %steal %idle
10:20:01 AM all 0.26 0.00 0.04 0.39 0.00 99.31
10:30:01 AM all 0.23 0.00 0.10 0.08 0.00 99.58
10:40:01 AM all 0.32 0.00 0.05 0.17 0.00 99.47
Tags: None
Subject
Written By
Posted
Oracle 5.6 Vs Percona 5.6: High CPU Utilization in Percona
May 15, 2015 01:47PM
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.