Description of problem: As volumes get fuller performance starts to degrade. I was adding data to a 20 TB volume and I started to notice the perf hit when it hit 2 TB(10% full). To quantify we created an empty volume on the same bricks and ran the following reproducer: # time tar xf boost_1_57_0.tar.gz On the fuller / aged volume it took significantly longer to untar. Version-Release number of selected component (if applicable): glusterfs-3.6.0.42-1.el6rhs.x86_64 How reproducible: Every time I tried so far. Steps to Reproduce: 1. Create teo volumes on the same bricks 2. Fill 1 up with data(~50%) 3. Untar the files Actual results: Performance degrades when the volume hits ~10% usage. Expected results: Performance does not degrade as the volume gets fuller. Additional info: We have this reproed in the lab, ping me for access to the setuo.
@Pranith, request you post your current analysis here on the same, thanks.
Thank you for your report. This bug is filed against a component for which no further new development is being undertaken. If you find the bug relevant in the higher versions too, please feel free to reopen the bug.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days