[Tech Request]: logservice propose Wait Time Significantly Longer than fsync Wait Time

sukki37 commented 2 months ago

Is there an existing issue for the same tech request?

[X] I have checked the existing issues.

Does this tech request not affect user experience?

[X] This tech request doesn't affect user experience.

What would you like to be added ?

During testing of the insert on non-primary key tables(standalone), we observed that the logservice propose operation has a significantly longer wait time compared to the fsync operation. The discrepancy between these two operations seems to indicate that the propose process, which is supposed to commit logs across replicas, is being delayed by factors that are not related to the disk I/O or fsync operations.

Investigate the root cause of the prolonged logservice propose wait time.
Explore potential optimizations, such as network enhancements, load balancing across replicas, or optimizing the consensus algorithm to reduce the propose wait time.

Why is this needed ?

This could be causing delays in transaction processing and impacting overall performance, especially in high-concurrency scenarios.

Additional information

out

tool: https://github.com/iovisor/bcc https://github.com/brendangregg/FlameGraph

command: /bcc/tools/offcputime.py -df -p "$(pgrep -d ',' test)" 30 > out.stacks ./FlameGraph/flamegraph.pl --color=io --title="Off-CPU Time Flame Graph" --countname=us < out.stacks > out.svg

volgariver6 commented 2 months ago

未处理

volgariver6 commented 2 months ago

未处理