Hi all, I have an issue where whether I am doing a SMB file transfer, streaming from Unraid to a media playback device via SMB, or even running a speed test via the OpenSpeed test, network traffic drops to 0% briefly and recovers every 40 seconds (ruling out that it is SMB related).

 

This is an example of when I was doing a large single file transfer to illustrate the issue. The main reason I am chasing this down is because my media players have small buffers so these drops cause the media playback to stall even at much lower data rates.

 

Things I have tried:

• Moved my docker.img off of the spinning disk array to my SSD cache
• Disabled the Docker service entirely to avoid any extra R/Ws being done to the disks
• Disabled NIC Offload in the Tips and Tweaks Plugin
• Adjusted Disk Cache settings to the below but also tried various values
  • vm.dirty_background_ratio at 2%
  • vm.dirty_ratio to 4%
• Transferring data from various different physical disks including off of cache SSD

 

Hardware info:

• i9-9900K with 64GB of RAM
• LSI 9211-8i SAS controller
• 7x 18TB WD Red Pros (all SMART tests pass and the drives are less than a year old)
  • 2 drives are parity
• QNAP QXG-10G1T 10GbE NIC w/ Marvell AQtion AQC107 controller (used in previous NAS which did not exhibit this issue)
• Main PC and Unraid server on the same Ubiquiti Flex-XG 10GbE switch

 

If anyone has any ideas of other things to try, I am all ears and greatly appreciate any potential leads!

Does it show the same issue if you use the onboard NIC?

10 hours ago, JorgeB said:

Does it show the same issue if you use the onboard NIC?

That is a great call and will definitely be testing that out this weekend once I have time, thank you! My current NIC was fine in my old NAS but that certainly does not rule out that there is some strange interaction between the NIC and Unraid or even that it is beginning to fail.