xrichnj

this would absolutely work, but i'm not willing to do it because i'm suspicious that libssl1 has something to do with the recent vulnerabilities patched by LimeTech. they said if youre on 6.12.4 and before (which only has libssl1), then youre vulnerable. if you are on 7.0.0 (which ONLY has libssl3), then youre safe from the vulnerabilities. maybe someone can check what ssl libraries 6.12.15 has (which is also safe) to shed some light on this. look, i'm not a security researcher, but the above worries me about installing old and "unofficial" libraries. libssl1 is pretty antiquated at this point, maybe the discovery of these vulnerabilities is what led to it being dropped from 7.0.0. i'm sure Lime had a good reason for removing/replacing it, i dont think they would remove old libraries that would break things "because they felt like it", there needed to be good reason. having better fan control isnt worth leaving my server potentially vulnerable to xss and csrf

ah, ok i see. my drives are right upfront as well (its a rackmount server), so drive temps are never an issue with 6 fans pulling fresh ambient air across them always, i just need the cpus to respond to the cpu. i set the "low temperature threshold" value to above my idle cpu temperatures and tuned the "fan speed minimum" to the value i wanted the fans to run at idle and it has stopped the up and down ramping. ex: the monitored cpu idles between 44-47c. so i set the "low temp threshold" to 50c and set the fan speed minimum to 18.7%, and now it can bounce around between 44 and 47c without changing. it only starts applying the curve once the monitored cpu hits 50c. that i can live with, but the lack of dual-cpu monitoring options still concerns me. i have 28cores/56threads and dont like to have everything just blast away at whatever cores it can get its hands on, so i have containers and vms segmented/compartmentalized to specific cores or even just give it a whole cpu (but not both). the fact i cant tell the plugin to "use the higher of the 2 temps of the 2 cpus to control the fans" means i have to be careful what is running on which cpu to avoid the fans not increasing and thermals going out of control. i suppose i could monitor the exhaust temp and run the fans off that, as the exhaust temp correlates to either cpu being loaded, but that is certainly not elegant/ideal either.

under temperature sensor, your photo lists 'hdd temperature'. doesnt this mean that your fans are reading your hdd temp for control and not your cpu?

also, can you explain the 'standard' vs 'override' fields?

yeah they seem helpful in this thread, and receptive to ideas for improvement. i hope with some development, this tool can be great for fan control.

yes it 'works' for me too, but as far as i can tell, lacks an algorithm to select how to monitor dual cpus for fan control (you can monitor one cpu or the other, not the highest temp of the 2 cpus or the average between the 2 cpus). if im using the non-monitored cpu for something heavily, the monitored temp doesn't really change, the fans dont ramp up, and the thermals run away. also, there is very little granularity for customizing the fan curve itself. if your cpu is at idle and goes back and forth between 45 and 46 degrees, i havent found settings that will get it to stop changing the fan speeds with these tiny fluctuations in temp.

with this plugin? or you have a script that works with 7.0.0 for fan control?

yeah i thought so too when i upgraded, but havent found anything that supports the correct libraries, either. i'm using a dell r730xd

same, my script is dead since upgrading to 7.0.0. even with a symlink to libcrypto.so.1.1, it still requires openssl1.1, which unraid doesnt include with 7.0.0. been using this plugin as a stop-gap until i found a solution, but it hasnt come and doesnt look like it will with nerdtools being dead, honestly. any chance for some love for us dual-socket users with this plugin? an algorithm to select which processor to read the temps for fan control would be hugely useful. (average of the 2 cpu temps, highest of the 2 cpu temps). currently i can just select one cpu or the other to monitor. if the non-monitored cpu gets used heavily, the fans wont ramp up

this was it, i forgot it was even custom i was on 4.x.x for so long. new (stock) theme for 5.x.x kinda sucks, will look for something better designed for v5. thank you for the quick response and for all the hard work you put in. it is appreciated, @binhex

came back 3 months and a couple of 'point' update releases later and still the same result. what am i missing?

having an issue with torrents adding as "unknown" since the 5.0.0 update. they download when force resumed, but i lose control over them (cant pause, regular resume, or set speed limits). rolling back to 4.6.7-1-01 works fine as it always did.

worked for me today too!

@binhex, i have experienced similar on proton using wireguard since one of the last updates. the port does set right, but i no longer get the "successfully assigned and bound incoming port XXXXX" message, or the 45 second watchdog update in log to make sure its working properly. i like to check it every so often to make sure its still operational and injecting the port, should it change. my log stops any showing anything new past "[info]qbittorrent process listening on port" as @Sak mentioned above. the last time i saw these message properly was 06-03-2024 according to supervisord.log. did an update change this? or did i mess something up? thanks!

working as it should after 2024.03.09. thanks again!

somewhere in the last week between update 2024.03.06 to now (just installed 2024.03.08), it broke my "locate" function. it flashes in the GUI as it should, but i get no blinking on the physical tray itself. i deleted all old databases and started a new one (force update all), reconfigured the locations and colors and such, and no blinking trays. it has been a long time since i setup the plugin initially cant remember how i got it to work. what am i missing? thank you for all the hard work and constant updates!

i'm trying to configure a windows10 vm that has internet access and can access unraid SMB shares. my system has an intel quad-NIC and i would like to use eth0 for unraid administration/dockers/etc, and eth1 for the vm. the issue is when i try a long file transfer (either from the vdisk on the pool to the SMB share or vice versa), the speed fluctuates from full (112MB/s) down to 0 and back up to 112 over and over again. during this, dmesg shows [32318.935665] device eth2 entered promiscuous mode [32394.938876] device eth2 left promiscuous mode [32414.949728] br1: port 2(vnet0) entered blocking state [32414.949738] br1: port 2(vnet0) entered disabled state [32414.949900] device vnet0 entered promiscuous mode [32414.950377] br1: port 2(vnet0) entered blocking state [32414.950382] br1: port 2(vnet0) entered forwarding state i have tried about a million network configs and am getting the same results. i thought maybe the nic was overheating, but any other computer, windows or linux, on my network that i try the same file transfer from has perfect performance, so i'm thinking its something in the networking of the vm that is causing the dropping. i'm new to unraid, but not new to windows, smb, linux, VMs, or networking, but i am no netadmin, either. can somebody please sort me out here or at least get me going in the right direction? i'm pulling my hair out, ive never had such issues trying to transfer files to/from a network share, especially when the drives are inside the same machine!