Niklas

Looking good. Now I see the "UNRAID"-logo. Thanks!

Way too big Unraid logo in the e-mail notifications from the forums. More than half the mail is logo only before you see the important stuff, the text in the mail. ;)

I added my scripts to this FR (works on 7.3): https://product.unraid.net/p/optimize-boot-backup-by-disabling-compression-or-adding-a-toggle Quite easy to add notifications if you'd wan't that.

Official plugin spotted. ;) You can find it in Apps now. "Installs the apprise-go CLI as /usr/bin/apprise and adds an Apprise target under Settings > Notifications > Notification Agents."

Looking good! The freq stops at the last value it reads under load but that's what qmassa say so..

When using intel_gpu_top + qmassa the "GPU Clock (MHz)" show 0 when the card has no work. Only showing values over 0 under load. When using qmassa only it shows a frequency at idle, and thats probably correct. In my case it will never be 0 as my hw is too old to get the card down in the lowest power saving state. Could FRQ-GT0 from qmassa be used in intel_gpu_top + qmassa mode too? Right now qmassa show FRQ-GT0 1000/2450 (this is probably correct, 2450 is my card max) and intel-gpu-top show 0/ 0 MHz. In qmassa only mode it seem to use FRQ-GT0 but then I lose other stats ofc. Edit: Under load, intel_gpu_top shows 864/1279 MHz with both values jumping around, qmassa looks correct with ~2100/2450 (snapshotted just now), only the first number changes. Like 2100 of max 2450. My frequency will never be 0. At the same time qmassa has FRQ-GT0 and also requested and actual. Unraid 7.3.0-rcs qmassa: FRQ-GT0: 800/2450 Requested [767] Actual [800] intel_gpu_top: 0/ 0 MHz /sys/kernel/debug/dri/0000:03:00.0# cat i915_frequency_info PM MASK=0x00000000 pm_intrmsk_mbz: 0x00000000 RPSTAT1: 0x00018103 RPNSWREQ: 767MHz Lowest (RPN) frequency: 300MHz Nominal (RP1) frequency: 600MHz Max non-overclocked (RP0) frequency: 2450MHz Current freq: 767 MHz Actual freq: 800 MHz Min freq: 300 MHz Boost freq: 2450 MHz Max freq: 2450 MHz efficient (RPe) frequency: 600 MHz But the frequency seems to stop at what it was at when the load was removed so idk. ;) Yeah, it stays on the last reading when it goes back to idle so the freq is probably not updating as it should with no load..

Working! Very nice. Thanks! It could be mentioned that the numbers report "rpm" but that's not important. Edit: This is what I said to qwen3.5:0.8b to get the fan to turn on. Lots of thinking there. Hejsan means like a friendly "Hi" in Swedish. ;)

Thanks @SimonF! PCIe Gen (Max): 3 (4)Lanes (Max): 8 (8), looking good. Just a missing space/delimiter before Lanes. Fan still shows N/A at 0.

Looking good! PCI bus utilization shows more (not stuck on 1 as usual with Linux and Arc). Fan show N/A when qmassa show 0. When it spins, it shows up but then back to N/A when it stops. Provoked the fan to turn on with some local ollama ai work. The ?-active app is qmassa. GPU Clock (MHz) show 0 when not under load but that's normal maybe?

0 is probably correct right now. Also, vram looks correct. It also shows power. Power draw in the plugin does not work (always 0 and has not worked before but looks to show there?) What I see:

Memory usage works with the xe-driver (but some containers does not like xe so back to i915, maybe in the future). Fan shows N/A on both. Worked on i915 on old release.

Is this my usb flash drive "spinning down"? emhttpd: spinning down /dev/sdk emhttpd: sdspin /dev/sdk down: 1 Does it save power/wear? 😁 Using USB for license (my hardware is too old to have or add TPM). sdk = usb drive, old boot.

I usually use dev tools (F12) and do empty cache with hard reload. Did it now and at least the memory usage bar fill went away, fan and memory usage N/A but memory usage may not be available on Arc A380 but fan is. Edit: The fan maybe needs to turn on at least once for it to not show N/A? I'll keep my eyes on it.

After the update, my fan show N/A instead of 0 (it almost never spins up). Memory usage I think is new, at least that bar.

It's a bit cpu intensive, yes?

No. We had beta 2 and now https://forums.unraid.net/topic/198402-unraid-730-rc1-now-available

We on rc.1 now?

Select "Unraid OS Releases (Next Branch)" as this is a release candidate.

My server did that from time to time when rebooting. Just showing a blinking cursor after post when the boot menu was supposed to load. Had to push reset when that happened. Haven't seen that happening with internal boot.

Nice! Up and running with the random mac-problem solved (I have added fixed macs to my containers).

So this thing with random mac's generated at docker or server restart is present in this version too I guess. https://product.unraid.net/p/7-3-0-beta-1-containers-start-with-random-mac In my case I added --mac-address to all my containers but then I got hit with extremely slow docker service start as containers fully rebuilds every time with --mac-address present. https://product.unraid.net/p/docker-containers-extrememly-slow-to-start So I had to remove my --mac-address's but now my router is flooded with random macs that I manually have to remove from the offline list (you could ask how often i reboot but it does happen and every time it adds ~25 new macs as offline). Updating a container generates a new random mac and the router see traffic from the same ip but also the same ip used by several macs (I set "Fixed IP address" for my containers) creating a mess. I mention this here because I got pm:ed by another user looking for a solution as this also happens in 7.2.5-rc.1 but I'm on 7.3.0 beta. I understand this has changed in later docker versions. I don't know enough about how docker works to fix it myself. Before it was fine with the generated macs as docker remembered and reused them but now it's some type of random every time-thing. ;)

I have a couple of these: /mnt/user/appdata/app123 → /config pool-only share · Share appdata is pool-only on ssd, but the app still goes through /mnt/user. This is not a problem anymore (for me at least, I always use /mnt/user) with exclusive access/shares (6.12+).

That ai-answer mentioned RST. Not relevant for Unraid (Intel Rapid Storage technology is Windows only and you don't need to activate it?) Inconsistent behavior? As in what? The conclusion reads more like faster options exists and Optane is "obsolete". Optane is cheap, absolutely fast enough, and reliable for this kind of use.

Possible to get mptcp support added to the kernel (could be used by proxies like npmplus)?

I know what it is, doh! It is when I have --mac-address="02:42:xx:xx:xx:xx" set in extra parameters. It rebuilds every container on every start of docker if that is set. It takes time and for big containers it can delay start for minutes. Removed all my set mac's and now back to normal (but router sees random macs for my containers again). Back to internal boot.. 😅