Anyone Has Found Resolutions For Cpu_iowait or Is It a False Negative From Glances?

[ButchR]

I have not been  successful  in finding  any resolution for my Unraid  cpu_iowait issues.  There are others with this issue that have not found any resolution.    Please share any good successful comments with me on this issue.  

[ButchR]

Anyone having these issues?

 

[Squid]

IO Wait isn't an "issue" the way that I *think* you're thinking that it is.

 

It's a metric that reflects that process(es) are waiting for I/O to complete from the drives / ssd's.  This could be because of multiple reasons -> the drives simply can't keep up with the request, too many simultaneous requests happening, failing drives, drives continually dropping off and reconnecting, exceeding the bandwidth by a fair margin of what the PCIe bus can handle etc.

 

There unfortunately is no one size fits all solution to what boils down to be a rather generic term.

 

As netdata words it's comments on the metric:

Keep an eye on iowait  (0.17%). If it is constantly high, your disks are a bottleneck and they slow your system down.

 

All that being said, from the screen shot you've included, the process Plex Transcoder is writing to disk3.  Assuming you have a parity drive, that's going to be a major bottleneck, as writes to a parity protected array are inherently slow, and the storage of the temp files for transcoding belongs on an SSD.

 

Or, put another way: Nothing is stopping you from using a DVD-RAM drive (I'm one of the few people who owned one) as a cache drive.  You're I/O wait is going to be sky-high.  But nothing is actually wrong because it's just a slow device.

[pinion]

Can you tell me how you determined what process was causing the iowait and how you determined the disk it was using? I have the same issue I'd like to figure out where the bottleneck is and I'm struggling to find what you see in the screenshots.