dazzathewiz

Thanks very much - I will order another drive to replace that one and then play with those tests. The docker is great - the reason I used it was because I am seeing very high CPU IO Wait times when copy operations to the Array are happening and I couldn't pin down the cause. I suspect the issue will be this drive. So thanks for your efforts developing this one.

So ah, what would be the advice given this result... 😰

I was having the same problem, noticed it updated over night and now doesn't work. TL;DR this happens in the new version because the docker no longer runs under root and I had some extra parameters that's running `apk` which isn't allowed by the new telegraf user. Longer explain: A google lead me to info about it being a user permission issue: https://stackoverflow.com/questions/50727783/alpine-docker-error-unable-to-lock-database-permission-denied-error-failed-to Then looked at the dockerfile in github, it seems prior to Oct 15, the telegraf docker was running as root. They changed this to use 'telegraf' user, see commit here: https://github.com/influxdata/influxdata-docker/commit/83520188be87cbee6edb3ac995c647cf99acdf51 The error in my case was happening becuase I add some apk pakages by docker settings -> advanced -> Post Arguments /bin/sh -c 'apk update && apk add lm_sensors lm-sensors-detect perl && apk add smartmontools && apk add nvme-cli && telegraf' After removing those argument, I simply get an error in the log because it can't find smart (which was removed): After commenting out inputs.smart in my config I could start telegraf again. In my case, the graphs and data I use actually wasn't affected by not having inputs.smart, so I didn't reconfigure it. But this may help others if they don't want to pin telegraf to an older/specific tag.

Oh awesome, thanks! Can confirm adding options i915 force_probe=4c8a to /boot/config/modprobe.d/i915.conf worked a treat and I now have the /dev/dri device

Have an i7-11700 and same issue. It doesn't look like it's updated in the Kernel for support yet. Just have to wait - I have tried 6.9.2 and still not updated.

Yay!!

If I can just 'pile on' to this thread... I have a MB with integrated Bluetooth (uses Intel 8260 communication module, which is connected via PCIE) and I want to pass that through to a docker. If I was using a VM instead, I'm sure I could just pass this IOMMU group: IOMMU group 17:[8086:24f3] 06:00.0 Network controller: Intel Corporation Wireless 8260 (rev 3a) But I would prefer to use the docker so would need the kernel module. Is the correct module called 'btintel'? I guess it's not supported in the current kernel given: modprobe btintel modprobe: FATAL: Module btintel not found in directory /lib/modules/4.18.17-unRAID

I've looked into this myself in recent months. TLDR; No but personally I have a custom script which works for my setup, but there some inconvenience in setting it up and maintaining it... I also noticed someone has recently made a request for this in telegraf (https://github.com/influxdata/telegraf/issues/4169) If you put this code into a script called nvmetemp.sh in your /mnt/user/appdata/telegraf: #!/bin/sh for device in /dev/nvme[0-9]; do temp=`smartctl -a $device | grep Temperature | awk '{print $2}'` serial=`smartctl -a $device | grep Serial | awk '{print $3}'` percent=`smartctl -a $device | grep Percentage | awk '{print $3}' | sed 's/.$//'` echo "disk,device=$device,serial=$serial temperature=$temp,percent_used=$percent" done Then you need to change a couple of things inside the docker (which you get a shell to by typing "docker exec -it telegraf /bin/sh" without quotes): (**Note comments about adding smartmontools package below) apk add smartmontools chmod 755 /config/nvmetemp.sh **Note smartmontools won't be installed the next time your docker updates to a new version. So if you have auto-update dockers it disappears and your temp is no longer recorded. You have to do the same process of adding the package inside the docker (apk add smartmontools) every time the docker updates. Then un-comment in your telegraf.conf [[inputs.exe]] so it should look something like this: [[inputs.exec]] commands = ["sh /config/nvmetemp.sh"] timeout = "10s" data_format = "influx" The actual problem is the method used to get disk temperatures by Telegraf - I think it grep's the string from the smartctl -a /dev/*** command output. NVME devices have a different text format in smartctl, thus Telegraf doesn't find the NVME temp. It seems to me like the plugins [[inputs.sensors]] and [[inputs.smart]] use the same method of getting the disk temperatures.