That works fine. The problem downloading Bitcoin logs concerns me.
Well, it works for now. I don’t know how long it will be like this, but I’ll keep an eye on it. Thank you for your patience
I don’t see an obvious cause in the OS or kernel log.
nope… same, but slower
200-250 blocks and reboot and again on specific block like 914563 now
was fun, gonna leave this for few days. thanks
Did you swap the RAM, and do it again?
yes, did this with same stick - different slots, and put back two sticks and one from another pc (in this case it didn’t start) but same result.. don’t know, maybe ssd.. but both stoped almost on same block?
IBD 100% – successful.
The issue with RAM during synchronization remains, but otherwise it works now properly and without complaints. Satisfied, thank you.
Okay, cool! So how much good RAM do you have in the geekom a5 now? What do you intend to do with the Nuc? Get some new RAM for it?
GEEKOM A5 - I put back the two 8GB sticks that came from the factory – it works fine. I ran Elect Datum Mempool LND , it’s been running for two days without any issues, system load is minimal.
(Again… from the start, to block 912000 everything was fine. Then, on the 8GB stick, I gradually loaded everything else – there were reboots from time to time, but less often (sure, because of the speed) and from time to time I was stopping the service so it could save the work, if I can say that, cause its downloaded like 200 blocks and rebooted , I try again, download 180.. stop the service, start again.. work for hour and half.. rebooted )
I’m using the NUC for Pi-hole for now, not sure about later (or maybe, I would sync on external drive then copy in to nuc using ssh)
I think I’ll buy a NUC 10–13 or a Genmachine for the next one.
Thanks again, so far everything works and I’m happy with it..
It shouldn’t be acting like that, and it’s likely to get worse, rather than better. I’d be concerned about the reliability of your hardware. It’s likely to be an ongoing source of frustration.
1 Like
I completely agree… some element of the system turned out to have a problem, or several at once. For now, I’ll see how it works, then when possible, I’ll try to eliminate one problem at a time. I think I need to try getting a 2–4 TB NVMe for the system, then RAM.
Yeah… you were right. After the power was turned off for two hours, the A5 turned on and started working properly, but then it began rebooting after downloading every 10–15 blocks that he missed while was off. Then it happened again, and I had to leave only 8 GB (short-term solution).
Intel NUC 8 with NVME by the way — nothing changed, the same issues, corrupted block…
I think we can close this thread)
Okay, so just to be clear, the conclusion is that both machines have bad RAM?
If only I knew..
The problem is only with Bitcoin.
In other scenarios, everything is fine, and no errors occur, no matter how I use these mini PCs.
windows + brute force on the CPU at a temperature of 80°C — it holds up.
Jellyfin plays a movie to an external device via the app well … don’t know how can i load ram to test it in working scenario, meatiest 86 - errors:0
** .. its ok that if I change data cache from blank or 12288 to 6144 it doesn’t decrease speed with two sticks of ram?
So you ran memtest, and no errors?
The IBD is the most resource intensive thing your server will likely ever do. As Stuart pointed out above, it will bring into the light hardware issues that otherwise might go unnoticed.
Once the IBD is complete, you should blank out the database cache field, to set it back to the very small default.
yes, memtest with two sticks, one time with one.. no errors.
indeed, IBD is a great stress test.. mine didn’t passed. But still, from 0-100% yes, I agree .. but last 15 blocks and same issue, strange.
ok..thanx, ill think about it. Thank you for your time, sorry for taking up so much of it.
You might consider testing both sticks together. You’re not using the dual channel feature with only one of them. It’s unlikely that bad ram would pass memtest. So the SSD’s are your next prime suspect.
I had ram pass memtest so I thought they were good. I went through weeks of troubleshooting uggh! I later retested the same ram on Lenovo diagnostics and it failed quickly.
Hope that helps somehow & good luck
Interesting. I’ve never seen bad ram pass memtest. Yes, don’t use ram that fails a test anywhere. Corruption and data loss will likely result. Not to mention all this frustration.