SageTV/OpenDCT causing extremely high disk IO

[Korny]

I've been running SageTV for about 3 weeks in a unraid docker along with opendct. I've been having intermitent issues with extremely high disk IO spikes to the point that my virtual machine died, and even some dockers didn't survive. I caught one a couple days ago where is spiked my user load to 120-130 and a wait time of 90+ . It thought it was my cache drive dying so I replaced with with a 256gig M2 drive over the weekend. On Sunday I caught it again with 6-8 "threads" in iotop at 99 utilization. I want to stay with SageTV, but if this continues I may have to explore other options. I disabled SageTV dockers and OpenDCT and am running another PVR in a virtual machine and have not experienced any IO issues since. I've attached my unraid diagnositics.
Any help is appreciated
Thanks
Aaron

[EnterNoEscape]

Can you also post the SageTV and OpenDCT logs?

[Korny]

I'm running SageTV and OpenDCT right now to see if I can get the condition to reappear. I will post them if/when I can reproduce it. Hey EnterNoEscape, I have the same motherboard/cpu you do. Have you had any issues using C-States? I had it set as auto in the bios and it was causing hard locks.

[EnterNoEscape]

Originally Posted by Korny

I'm running SageTV and OpenDCT right now to see if I can get the condition to reappear. I will post them if/when I can reproduce it. Hey EnterNoEscape, I have the same motherboard/cpu you do. Have you had any issues using C-States? I had it set as auto in the bios and it was causing hard locks.

You've really got me thinking now because the last real problem I had was a while ago. I believe I had a problem with them, but the real issue was a bad memory stick.

[Korny]

Originally Posted by EnterNoEscape

You've really got me thinking now because the last real problem I had was a while ago. I believe I had a problem with them, but the real issue was a bad memory stick.

I just replaced 2 memory sticks a few months ago, and ran memtest86 for 2 days without issues on the new ram + existing sticks. I'm not complaining to much, I saw an idle wattage of 140 this morning with 14 spinners, a VM running and multiple dockers.

[Korny]

Well.... had an old disk die today, never threw a smart error. We will see if that was the cause. Although I'm not sure how it would have caused it since SageTV/OpenDCT were writing to the cache drive, but maybe it caused a hang in an IO process that slowed everything else down.