Worker names are hashed and passwords dropped at capture time.
Share this in the Bitcoin section of the forum: Bitcoin - PaulsCode (a free account is needed to post)
or, if you already use GitHub:
I have no idea about the technical details, because Iâm mining for the first time; thanks a lot for the great tutorial on YouTube. My general impression is that itâs workingâwith issues, donât know whether these are HW-problems or the SW could be improved: The figures are rising, which looks normalâBUT the number of the HW errors displayed is higher as the number of âacceptedâ shares. After more than one hour these two numbers are about the same order of magnitude.
The HW-Error in this case is not a problem The miner is just interrupting itself, because of the very low difficulty adjustment on regtest when it starts. Basically, your ASIC is finding a block every five seconds or so, and as soon as it confirms, it obsoletes every share that was in progress on the duplicate block. And anything the chip returns against a job that no longer exists canât become a share, and it appears the firmware just counts those jobs as HW-Error.
We can probably confirm my analysis though by running on Testnet which appears to have difficulty-adjusted down considerably. Connorâs Lab released a video on his setup HERE that might be worth replicating if you have the hardware (looks like he set up a dedicated Umbrel node in his case). I probably should look at simplfying his process and supporting Start9 as well as installing alongside existing Knots and Datum setups (may already be possible, I havenât looked into it yet). Iâll look into it later when I have some time if he or someone else hasnât already released something.
Thank You very much, I bought this miner used and I will try to run it from early on and we will see. I watched Connorâs Lab, however I donât want to disturb my Raspberry-Umbrel-Knots-Node running. I work on Start9, I need advice and will look what I can find to get this SC-Box running on testnet and later on the real net. Your advice would be very much appreciated, I am not able (yet) to download and install SW from original GitHub pages, I need it in the Start9 download folders.
For information, the miner has a /#/debug page, and there I found the following pic with overviews of the âchipsâ, everything looks normal with the exception of the HW-Errors, the numbers indicated fit to the total number of errors shown on the homepage. This looks like a possible HW-problem, but could also be certain âtimingsâ, which lead to the errors culminating on three of the chips:
Interesting. That counters my earlier theory about the difficulty being too low. Weâll know for sure when you start mining on a higher difficulty (if the HW errors continue, then constant re-orging isnât the cause). Iâm leaning to it more likely being something physical specific to your device (such as old thermal paste, a literal bug stuck in a heat sink, or just failing old chips for that matter). Timing is also a possibility (though that wouldnât be specific to Bitcoin I donât think, maybe a fine-tuning issue)
You could probably narrow it down further by using a Sia address (the one below is a donation address for a random Sia contributer I found) and mining for a few minutes to a pool that supports the coin. For example:
Pool URL:
stratum+tcp://sc.f2pool.com:7788
Worker:
26117c19ca3975b315d663dcbbc19cf9c07274f441689d4392ed380b2337589ef1aacfbdc93f.SCBoxII
Password:
x
To see if the HW Errors go away (definitely not a hardware problem) or are still present (hardware problem likely)
Got it. I think the next test will be to see if low difficulty is a factor. Iâll put together a simple method for joining the Testnet (if Connorâs Lab or someone else doesnât beat me to it).
For information, I ran Knots and Datum with Umbrel on the testchain (according Connorâs Lab âBitcoin Blake2b PoW Testnetâ). This miner works with the current SW-versions and I could create several blocks on the testnet, although the âHW-Errorâ appeared again with about double the number as the number of âacceptedâ.
The latest releases of the two apps allow switching to testnet4 (Knots Blake2b > Actions & Config > Select Chain). You will need to wait for Knots to complete the IBD on testnet4 once you switch to it (it is a lot faster than mainnet), but once it finishes you should be able to mine on testnet4. Iâm still working on supporting pruned nodes with electrs + Sparrow and Mempool updates (which will be the subject of my next video). But I thought Iâd let you know about the update earlier, in case you want to check if running on testnet4 still has those HWErr numbers.
For your information: I was able to reduce the error rate somewhat by using a power supply that allows for fine adjustment of the output voltage. I am mining in full hash mode at 12.5 volts, and HW errors are staying below 50% of the âAcceptedâ count. I donât know if this is a coincidence or a fluke, but ultimately, there is also some voltage drop along the supply line. The operating temperature remained unchanged (just under 70°C with additional fans placed nearby), and power consumption (including the PSU) also remained constant at 422 watts despite the voltage increase.
Questions:
Switching from Umbrel to StartOS and trying to mine, I have some âlittle questionsâ:
After having started Knots on StartOS, how can I recognize the status of the node IBD (highest block it has received)? Umbrel has a Dashboard showing connections and latest blocks. Here I have no glue how to check it without Mempool oder DATUM.
Starting DATUM, I missed to recognize and store the username (is it âadminâ?) and the password. Now I cannot open relevant pages of the programm. How can I get the password? I reinstalled DATUM, but the tag with the password information is not shown any more.
