alex3305

I have tested with a Pi 3B+ and ser2net which worked fine, but was IMHO very underpowered. So currently I have setup my old J4105 NUC-like with Debian, Docker and Docker Compose. There I just run Zigbee2MQTT and Z-WaveJS as a stack. I also added some other applications, such as a secondary backup server and secondary DNS. This setup consumes about 3W. Which is still less than the approximate 7W that the 13500 system consumes. This is probably an unrelated and different issue. ---- There also has been some discussion on the Intel Community regarding this issue. But still doesn't acknowledge it, but refers us back to 'Linux', whatever that means. Intel Community | High power consumption when using USB1.1 serial devices under Linux Intel Community | Poor power efficiency with B660/B760* boards and active** Full-Speed USB devices

Thank you so much for implementing this feature and making it work. You have really made my day! Edit: Whoops. It seems that collapsing folders when enabling this feature doesn't work anymore.

Thanks for the quick reply, but I kindly disagree with you that it would mess up the table alignment. The difference is that your HTML would be adjusted from: <tr> <td class="ct-name folder-name">...</td> <td class="updatecolumn folder-update">...</td> <td colspan="3">...</td> <td class="advanced folder-advanced">...</td> <td class="folder-autostart">...</td> </tr> to <tr> <td class="ct-name folder-name">...</td> <td colspan="4">...</td> <td class="advanced folder-advanced">...</td> <td class="folder-autostart">...</td> </tr> Than you'll get the same alignment as I have in my screenshot, without breaking anything. Therefore i would kindly ask you to reconsider or look into it ☺ Thanks! Insta-edit: I would do this with Custom CSS myself, but since there is not CSS equivalent to colspan, this seems pretty impossible to do so.

What a shame that labels aren't returned in the unRAID API. That would make life so much easier :/. But I have another feature request, if possible. I use Folderview (which is working great, thanks!) mostly for grouping my Compose applications that I've created outside of unRAID's DockerMan. However update reporting for those containers is wonky at best. And updates aren't working at all. Would it be possible to hide the Update Available message for those containers, and perhaps expand the folder? Something like this: Here is an JSON API excerpt of two of those containers: [ { "running":true, "paused":false, "autostart":false, "cpuset":"16,17", "url":"http://[IP]:[PORT:8123]/", "icon":"/state/plugins/dynamix.docker.manager/images/homeassistant-icon.png", "shell":"/bin/sh", "registry":null, "Support":null, "Project":null, "DonateLink":null, "ReadMe":null, "updated":"false", "template":false }, { "running":true, "paused":false, "autostart":true, "cpuset":"16,17,18,19", "url":"", "icon":"/state/plugins/dynamix.docker.manager/images/mosquitto-icon.png", "shell":"sh", "registry":"https://hub.docker.com/r/cmccambridge/mosquitto-unraid/", "Support":"https://github.com/cmccambridge/mosquitto-unraid/issues", "Project":"https://mosquitto.org", "DonateLink":"", "ReadMe":"", "updated":"true", "template":"/boot/config/plugins/dockerMan/templates-user/my-mosquitto.xml" } ] Thanks again for all the effort <3

Although hardly mentioned in this thread, I use cAdvisor with Prometheus for long term monitoring of my Docker containers. And because it is available in the Apps store I used that to install. However I noticed that the installed version is really outdated. The template seems to use the latest (default) tag for cAdvisor, but for some reason that points to version 0.38.6 on Google's Docker Registry. But the latest version is actually 0.47.2. This release is about 2.5 years newer than latest. This can easily be verified on the metrics page (i.e. http://unraid.lan:8080/metrics). This is latest: # HELP cadvisor_version_info A metric with a constant '1' value labeled by kernel version, OS version, docker version, cadvisor version & cadvisor revision. # TYPE cadvisor_version_info gauge cadvisor_version_info{cadvisorRevision="30515557",cadvisorVersion="v0.38.6",dockerVersion="20.10.24",kernelVersion="6.1.38-Unraid",osVersion="Alpine Linux v3.12"} 1 but adjusting the image to a certain version (i.e. gcr.io/cadvisor/cadvisor:v0.47.2), results in this: # HELP cadvisor_version_info A metric with a constant '1' value labeled by kernel version, OS version, docker version, cadvisor version & cadvisor revision. # TYPE cadvisor_version_info gauge cadvisor_version_info{cadvisorRevision="6730be15",cadvisorVersion="v0.47.2",dockerVersion="",kernelVersion="6.1.38-Unraid",osVersion="Alpine Linux v3.16"} 1 So maybe something the template dev can (easily) adjust.

Like I mentioned in my other reply, I already had isolated and pinned the VM to specific E-cores. But starting a VM - no matter what - prevents my Intel 13500 to enter any package C-state at all. So today I migrated all of my Home Assistant components to Docker containers. Mostly fairly trivial, except for some specific quirks. But running either Zigbee2MQTT or Z-Wave JS still prevents package C-state lower than C2. Power consumption is a bit better with about 25W in idle (running 45 containers), which was about around 30W with Home Assistant in a VM. But still nowhere near the 21W I was expecting.

Thanks for your input @mgutt, I really appreciate it. Currently I have pinned the VM to 4 of the 8 E-cores, but I also tried pinning it to a single P-core with CPU isolation, but that didn't make a difference. The host never gets to lower package C-states. Individual cores seem to get to C6 though, even on the assigned cores. Which makes me wonder if there is something off of my configuration. Btw I know that I can run HA and associated apps within Docker. And that would be my next test when I cannot get this working. I just kinda like the ease of use of HA OS.

Hi, I've recently migrated to a new system. A really similar system to this system by @mgutt. Except I've chosen the Intel Core i5-13500 and a be quiet! Pure Power 12 M 550W PSU. Thanks again, mgutt, for the inspiration and very good guidelines which were easy to follow. I'm trying to make this system (very) energy efficient, which I think is coming along quite nicely. Without anything loaded my idle usage is between 11-13w. Starting the array increases the usage to 14-15w. And starting my 20 or so Docker containers adds another 2-4w, totaling to about 16-19w. This is with 3 SSD's, 2 2.5" mechanical USB drives and 3 internal 3,5" drives. While I'm really pleased with the results, I noticed that when I start a VM (for instance Home Assistant) power consumption increases to around 24-26w. Starting up PowerTOP evidently shows why: and when I shut down the VM, C8 state is reached pretty much immediately: After some trial and error I'm at a loss why this happens and what I can possibly do about this. Except migrating away from the VM. I was wondering if anyone else has observed this behavior and maybe has a solution? Thanks!

That got lost in translation, I thought you meant a flash storage device such as an SSD. Anyway thanks. I've just added an 32GB pen drive and converted my array devices to pool devices and all seems to be working great. Thank you all for helping.

Would a flash drive work as an array device? Or only USB hard drives? I still have a couple of old 60GB Samsung SLC SSDs which I could potentially use. Or an old 120GB M.2 SATA SSD.

Great suggestion, but also no. I'm using the Gigabyte Brix J4105, which is really limited, but has been working well for me so far.

I only have space for 1 NVMe and 1 2.5" SATA disk in that system, so adding a small 'dummy' flash drive is impossible unfortunately. I read that adding an USB drive to the array would be possible. I will look into that. Thanks.

Ok, than I had that correct. But how would I run an optimal SSD only unRAID install than? Or is there a way to force TRIM on an array?

Hi all, I have been running unRAID for over a year now and I'm really satisfied with the product. I wanted to run unRAID as an application server with some Docker containers and a VM on my NAS and that seems to be working really well. To make the system silent, I have installed a 1TB NVME SSD in it from the beginning. And even with all the warnings about SSD's that seem to be working great. Recently however, I've added a second array device, another SSD. I don't really care about the data and parity, so I don't worry about that. I do worry about performance, and I read conflicting things about TRIM in the array. I'm wondering if this is supported now? I've been running my array in XFS mode as per default. Or should I perhaps run my installation differently? Because I'm open to suggestions! Thanks!