Description of problem: I was working on a .t where two requests are wound one after the other, the first fop that is wound takes more than a second(because of delay-gen) and the next doesn't. This was leading to both the fops taking delay-duration seconds instead of second fop unwinding earlier than the first one. When I debugged it, I found that io-threads is not matching threads to number of fops even when it can. The .t that I am working on for another bz will be updated at https://review.gluster.org/c/glusterfs/+/22674 Version-Release number of selected component (if applicable): How reproducible: Always. Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
REVIEW: https://review.gluster.org/22686 (performance/io-threads: Scale threads to match number of fops) posted (#2) for review on master by Pranith Kumar Karampuri
This bug is moved to https://github.com/gluster/glusterfs/issues/1065, and will be tracked there from now on. Visit GitHub issues URL for further details