Details
-
Bug
-
Resolution: Fixed
-
Critical
-
4.0.0
-
Security Level: Public
-
Yes
-
KV: Oct 25 - Nov 14
Description
Take a look at the attached screenshot of showfast. For version 3.1.0 and Sherlock 4.0.0-4019 RC2 , I have three data points each. You can view it if you click on them on showfast. The numbers are:
3.1.0-1776: 2.4, 2.5, 2.6
4.0.0-4019 RC2: 3.2, 3.2, and 3.7
which translate to an average increase of 34%.
The test (from my limited understanding) measures the warmup time of a bucket after it is restarted. This is a regression from 3.1.0 to Sherlock. Please investigate and fix.
Thanks!
Attachments
Issue Links
- relates to
-
MB-16169 [Windows] single KV node cluster seems to handle 100% set ops 7x slower in Sherlock
- Resolved