- 17.9K All Categories
- 3.4K Industry Applications
- 3.3K Intelligent Advisor
- 62 Insurance
- 536K On-Premises Infrastructure
- 138.2K Analytics Software
- 38.6K Application Development Software
- 5.7K Cloud Platform
- 109.4K Database Software
- 17.5K Enterprise Manager
- 8.8K Hardware
- 71.1K Infrastructure Software
- 105.2K Integration
- 41.6K Security Software
45 Second Periodic Performance Disruption
Has there been any resolution in the works to this problem or perhaps a workaround? If not I would be willing to help you debug & reproduce the problem.
- This irregularity is more prominent at higher queue depths
I'm running Centos 7.3 with kernel 4.9.1.
No core goes over 5% cpu usage.
The below graph is from a 4K Read, 32QD workload which gives me about 76K IOPs. This particular workload is run on a loopback device, but I have reproduced it on solid-state drives.
Here's my summary.html
As you can see, like clockwork there is a problem with the reported actual queue depth (and subsequently the IOPs) for a 1 second interval. This reported actual queue depth goes over (threads+1) followed by an interval that reports an actual queue depth of less than (threads-1) every 45 seconds.