-
Notifications
You must be signed in to change notification settings - Fork 2.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
MySQL 5.7.41 vs. 8.0.33 Performance Issue on amd64 Image #975
Comments
This is a really interesting and detailed summary, but I'm afraid I'm struggling to understand what we as the image packagers could do about it? 😅 We take the binaries published by Oracle (for both Debian and Oracle Linux-based images) and package them pretty faithfully with very minimal configuration differences from the stock packages provided by upstream. 🙇 |
I see what you are saying about taking the binaries published by Oracle. So, using the same specifications as defined in my "Environment 2" above, I installed the 2 different MySQL versions natively (eg. using apt-get install), then I ran the same Sysbench test. This was the results:
This is what the columns represent:
Based on these results:
So, I applied the same variable ("skip-log-bin") to the docker image for MySQL 8.0.33 running on a freshly installed AWS EC2 (t2.2xlarge) instance. The command is:
Then I ran the same Sysbench against this and here's the result:
The two columns called "5.7.41 (AWS EC2)" and "8.0.33 (AWS EC2 - vanilla)" were from my original testing, I just added them here for easier comparison. In the case of 8.0.33 with the skip-log-bin variable (last column), the performance is still not on par with 5.7.41. The query rate for 8.0.33 is 5533.94 queries/sec, while 5.7.41 is 7097.24 queries/sec. If we examine the native test, 8.0.33 with the skip-log-bin variable performs significantly better, reaching 10400.32 queries/sec, which is comparable to 5.7.42 9671.90 queries/sec. Given these results, I expected the performance of MySQL 8.0.33 with the skip-log-bin variable to have improved in a Docker environment, but it did not. Do you have some possible insights as to why this apparent performance change is not visible in a Docker environment? |
So MySQL will ways be slower on docker ? |
Issue Summary
When running a Sysbench test, it looks like the performance of MySQL 8.0.33 (amd64) is slower compared to MySQL 5.7.41 (amd64) when running in Mac (arm64) or in AWS EC2 Instance (amd64). I would have expected that 8.0.33 to be faster given that it is a newer version.
Environment
Steps to Reproduce (applicable for both environments)
Setup of the 2 MySQL versions
Run the tests
Sysbench Results on Mac Pro
For 5.7.41 on Mac Pro:
For 8.0.33 on Mac Pro:
Sysbench Results on AWS EC2
For 5.7.41 on AWS EC2:
For 8.0.33 on AWS EC2:
Summary of the Results
Expected Behavior
Based on the results, I expected that MySQL 8.0.33 should be faster given that it is a newer version, but it looks like MySQL 5.7.41 is faster. Is this a known issue?
The text was updated successfully, but these errors were encountered: