TexasUnraid

Pretty sure they will just stay where they are. I know it copies the plots and doesn't move them. I have it set to put the finished plots on the same drive in another folder but it still copies the data instead of moving it.

Yep, I bought a few things but they were already in my cart for the last 6 months and when I noticed prices skyrocketing and availability plummeting the other day realized I better buy them now or miss out all together. Take a a look at ebay sometime, supermicro 846 chassis that used to go for $200? Now the few that get listed go for $1500++ and they SELL! Almost enough for me to consider selling my backup chassis if I thought I could replace it. Everything I got so far expect the 146gb 10k drives was stuff I needed/wanted anyways, so worst case for me, I am out $15 lol

Yeah, I just started using daily linux last year myself. I won't be daily driving it anytime soon but for set it and forget it things like this it works well. Course I plan to use that GUI docker posted earlier to make things way easier. Right now I am just playing with chia, I see no point to go hardcore until pools are released and real world earning can be seen.

I would just ditch windows long term, Most report a 10-15% improvement in plotting speed on linux and I saw another test that compares NTFS, EXT4 and BTRFS and that should another 10-15% improvement moving to BTRFS IIRC.

I have seen a lot of reports that plotting is faster on linux and BTRFS is faster still then most other file systems. You might try running Ubuntu and see what the plot times do. The other thing is that the first stage is the only one that can handle mutiple threads, the rest are all single threaded, so single thread performance will play a big role there. I am running 2x 2670 CPU's which have a lot higher single thread performance then you, although surprised you can't get more as I am not even stressing it at this point, just doing basic tests until pooling comes online. I also increased my dirty writes in linux to 90%, so it can use up to 90% of my available memory as a cache for writes to the drives, didn't seem like this helped a ton but it might of had a larger effect then I thought. It reduced the thrashing of the drives though. based on my testing so far 3x parallel for a single drive netted me the best times per plot at 6.5 hours. If I was to do that on the 4x drives I am now testing in the raid, I would be around 2.2 hours per plot, a bit faster then the 2.5 hour raid0 performance of the drives thus far. Testing now with some staggering of the plots to see if that improves things. So yes, using the drives individually does seem to net slightly faster speeds all else being equal but it is a lot harder to keep track of, so there is a trade off. Watch the performance metrics of the system during plotting and you can find your bottlenecks. I use netdata on linux, allows me to log the whole process for later review. I don't max out the single drives IO until running 3x in parallel per drive (on these fairly fast ~170mb/s drives). With 2x plots there are times when the drive activity drops to zero. CPU limits the ability to plot a single plot any faster, thats why I need multiple in parallel to max the drives IO. I have 32 threads but they are ~7 year old threads, so they leave some performance on the table lol. It is looking like my limiting factor will end up being memory actually as I can only run ~17 in parallel with 64gb of ram but can do a lot more then that with my CPU's and disks before those max out. I won an auction for some cheap ECC DDR3 the other day, gonna see what 128gb of ram does with ~25-30 plots running in parallel. Yes I will be over-provisining the CPU but that is not an issue if I stagger the plots as only the first stage is multi-threaded. Plus worst case it slows down a bit.

Nice GUI docker, that is exactly what I have been looking for to start plotting on my main unraid server, might as well put it to use. Saw some cheap DDR3 ECC on ebay the other day, picked up some but thought about getting enough to make a ramdrive for plotting lol. Research showed it was not as good as it sounds though since the bottleneck actually lies in the CPU for the most part. Looking through my cart at some parts I have been putting off getting, if there is something you need/want for your server, get it quick. Prices are SKYROCKETING! I picked up an adaptec 24 port HBA 72416 a few months ago for $80 WITH 6x breakout cables (worth ~$90 by themselves). The same card is not going for $190 without the cables! I wanted to get another just for the cables but waited too long.

Thank you! I have been waiting for something like this to even try chia on unraid. Hopefully have some time to try it out this weekend after my current plots finish.

Yep, pretty impressed myself. Just finished the 12x plots and it was a tad slower then 8x on the 4x4tb raid0. Guessing if I staggered them I could get ~10-12 plots to be a bit faster with some fine tuning. As it is, just going to toss all my old drives that are basically useless to me now into a system and plot on those. Just got to make sure that you only raid together drives of similar speeds. I managed to get 20x 10krpm sas 150gb drives for $15 shipped on ebay the other day, so a bunch of drives I don't care about to play with and should net some decent performance.

Just a little FYI from some testing I am doing. You can get some good plotting speeds with normal hard drives without the need to wear out SSD's. With a regular 7200rpm 4tb single drive I was able to get an effective speed of 6.5 hours by plotting 3 in parallel. By raiding 4x 4tb drives together I have gotten it down to 2.5 effective hours/plot doing 8 in parallel and trying 12 now. I have a bunch of old small drives I plan on raiding together and using for plotting. Just a heads up so people don't think SSD's are required to plot. That said I would still use hard drives you don't care about for plotting as it will put a lot of wear and tear on them. I am doing the tests on an ubuntu machine using BTRFS filesystems. You can do the same thing with pools in unriad easier. I plan to do that once I settle on a setup.

Agreed^ Also being able to rename the sensors (fans in particular) would be very helpful. Along these same lines, is it possible to add multiple sensors to the fan control. Basically my mobo only has 1 PWM control. I can set the fans to be quite and everything is fine as long as it is not under load, if the drives start working the fans need to spin up, ok fine. The issue is that if the CPU's with passive heat sinks start working they also need the fans to spin up but right now I have to pick only one of those. It would be nice if it could have use multiple sensors for fan control and simply set the fans to whatever is higher. Thus allowing it to spin the fans up for either the CPU or Hard drives. Bonus points to add even more sensors as I could spin the fans up if the HBA or memory starting getting hot for example etc.

Yeah, that would work, I think there is just one file that needs to be replaced but it is simpler for me to just copy the whole folder over.

No, i just resorted to making a copy of the qbittorrent folder in appdata (named it -backup) and I just copy it over the existing qbittorrent folder and restart the docker. I should setup an rsync job so I don't loose any new seeds/downloads but have not been using it much lately so kinda dealing with it.

Just an observation, I have been having some random blackouts and the server randomly shutting down recently (UPS died the other day). Each time the server does an unclean shutdown, Qbittorrent seems to loose all it's settings. I now have a backup of the appdata folder and if I copy it back over the appdata and restart the docker, everything is back to normal. Any idea why the settings are erased on unclean shutdown?

there is a dos supermicro ipmi tool you can use that can change the ipmi ip and a bunch of other settings. The ip address is generally set in the bios. On my X9 board it has a setting to change the IP and it will use DHCP to get a ip and then set that as static. If you want a custom static address, my guess is that ipmi tool from supermicro.

also curious what this docker is for? https://hub.docker.com/r/ich777/debian-builder

Long time fan of your dockers. I have been slowly moving everything possible over to dockers on the server and need to make some custom dockers to complete of putting everything on the server lol. I found this dockerhub from you: https://hub.docker.com/r/ich777/novnc-baseimage Are there any details about how to use this? This sound like exactly what I need to setup some dockers for programs that don't exist yet. My immediate goal is to make a GUI docker for slic3r but the same should apply for a lot of other programs I think.

So just setting up my new-to-me server with a supermicro X9dax-7f mobo. It has built in power usage tracking via ipmi and I can get the output if I poll it manually: ~# ipmitool dcmi power reading Instantaneous power reading: 112 Watts Minimum during sampling period: 9 Watts Maximum during sampling period: 406 Watts Average power reading over sample period: 126 Watts IPMI timestamp: 04/16/2021 10:16:18 AM CDT Sampling period: 00078026 Seconds. Power reading state is: activated My question is can the Instantaneous power reading (or better yet multiple readings) be output to a sensor reading like the rest? Is there a way to add a manual sensor to poll this information? Ideally I would like to be able to see the power usage along the bottom taskbar but just having a place I could see it would be cool as well.

Here is what it says, looks like it has the motherboard voltages but not the CPU voltage / current. k10temp-pci-00c3 Adapter: PCI adapter CPU Temp: +44.5°C Tdie: +44.5°C nct6797-isa-0a20 Adapter: ISA adapter in0: 800.00 mV (min = +0.00 V, max = +1.74 V) in1: 1.02 V (min = +0.00 V, max = +0.00 V) ALARM in2: 3.36 V (min = +0.00 V, max = +0.00 V) ALARM in3: 3.31 V (min = +0.00 V, max = +0.00 V) ALARM in4: 1000.00 mV (min = +0.00 V, max = +0.00 V) ALARM in5: 152.00 mV (min = +0.00 V, max = +0.00 V) ALARM in6: 744.00 mV (min = +0.00 V, max = +0.00 V) ALARM in7: 3.34 V (min = +0.00 V, max = +0.00 V) ALARM in8: 3.30 V (min = +0.00 V, max = +0.00 V) ALARM in9: 1.84 V (min = +0.00 V, max = +0.00 V) ALARM in10: 0.00 V (min = +0.00 V, max = +0.00 V) in11: 712.00 mV (min = +0.00 V, max = +0.00 V) ALARM in12: 1.15 V (min = +0.00 V, max = +0.00 V) ALARM in13: 672.00 mV (min = +0.00 V, max = +0.00 V) ALARM in14: 1.53 V (min = +0.00 V, max = +0.00 V) ALARM Array Fan: 952 RPM (min = 0 RPM) Array Fan: 1288 RPM (min = 0 RPM) Array Fan: 1553 RPM (min = 0 RPM) Array Fan: 870 RPM (min = 0 RPM) SYSTIN: +40.0°C (high = +80.0°C, hyst = +75.0°C) sensor = CPU diode CPUTIN: +44.0°C (high = +112.0°C, hyst = +90.0°C) sensor = thermistor AUXTIN0: +40.0°C (high = +112.0°C, hyst = +90.0°C) sensor = thermistor AUXTIN1: -128.0°C sensor = thermistor AUXTIN2: +42.0°C sensor = thermistor AUXTIN3: -2.0°C sensor = thermistor SMBUSMASTER 0: +44.5°C intrusion0: ALARM intrusion1: ALARM beep_enable: disabled I used to see 2 separate sections in netdata, one would have the motherboard voltages and the other would have the CPU voltage and current. These are the same values as the motherboard readings. I could be wrong, just find it odd that the voltage and current disappeared at the same time that turbo boost stopped working. Honestly I plan to switch motherboards in the semi-near future anyways, mostly reporting this issue in case others have it. Although the random restarts earlier today were not comforting.

I didn't say it was part of unraid. I said it not displaying those values indicates that there is an issue with the driver as it worked before the update and I did not mess with the docker. Combined with the CPU not boosting to the correct speed backs up some kind of wider issue then just the governor changed. I don't know the in depth details but it pulls from the kernel /direct from hardware for a lot of things. Everything it displays can be seen from linux itself generally, it just puts it together with nice pretty graphs. If the driver is missing, then the graphs will not work, if the driver is installed they work. I am still new to linux so that is about the limit of my knowledge on the matter. I have also noticed that if I put a heavy load on the CPU the server restarted twice today which further makes me think it could be a driver issue of some sort. Although I was putting it under heavier load then normal so can't 100% say it was connected but it is out of the ordinary for sure.

It is a docker that monitors all of the sensors and readings from linux. The current and vcore used to show up in the beta but doesn't now which is generally due to a driver issue.

I don't think that the governor itself is the issue necessarily. Since the vcore and current readings from the CPU are missing as well, that would lead me to believe that it is actually some kind of driver issue. The vcore and currents should show up if all the drivers are working properly.

I have confirmed that netdata no longer shows the CPU voltage or current. So it would seem that some driver is not being loaded that is preventing all of these from working.

So, I just upgraded to 6.9.1 and found a bug. On 6.9 beta 30 that I was on before, setting the CPU governor to "Schedutil" would allow my ryzen 2600x to turbo to it's full ~4ghz speeds with on demand and performance limiting it to 3.6ghz. With the update to 6.9.1 Schedutil now also is limited to 3.6ghz and none of them appear to allow turbo past that. Also noticed in netdata I am no longer able to see vcore and CPU current for some reason, going to reboot and see if maybe it is just a driver loading issue.

Ok, cool, that explains that issue, I will try updating again at some point and doing that.

I updated awhile ago and it broke things, so I reverted to an older version. This got qbittorrent working but sonarr etc could not connect to it for some reason even though all the settings are the same. So I updated it again to see if things were fixed and first it is still not able to connect but in the logs it is trying to chown my entire media library that is stored under /data?? Why is it trying to change permissions on everything under /data? edit: reverting to binhex/arch-qbittorrentvpn:4.3.1-1-03 got it working again but if I clear the appdata folder, it will still try to chown my entire share mapped to /data.