I have a new Apollo BTC 2 unit I have added to my stable to compliment two Apollo BTC and one other Apollo BTC 2. This new one is proving to be a challenge to get running smoothly. It mines fine, but will not complete the initial blockchain sync. Twice it ran pretty quickly until 48% of the SSD was remaining as free space. I formatted the SSD both times it got 'stuck' and the second time I tried a new SD Card with a fresh image. Each time it had errors that increased in frequency with each attempt (timeout, connection refused, etc.) but I could restart the Node from the GUI, restart the unit, or sometimes had to power off and power on to recover. The third attempt to sync the blockchain took four days to get to the 48% disk space free and once again has stopped. The unit is now stable and not throwing out any error messages, but had not made any sync progress in the past 3 1/2 days, Still stuck with the current block of 770,203 (since Sunday morning).
At this point I suspect I have a bad SSD. All appears fine in Apollo Web OS. Show as a 1 tb drive. I have ordered a new SSD to try soon in a few days when it arrives.
Any other suggestions from anyone on what to try next would be welcomed. If it is not the SSD, then I am stumped, and reaching out to support directly has not been helpful.
First, have you checked your log file. you got a long way at 700K blocks. Your node may be stuck in a loop trying to confirm and looping through the last 100 to 200 blocks. It shows in the log.
I had similar issues and thought the same. The pcie is probably good, but you need a backup card anyway.
before using a new memory card You need to format the card. I had issues with the microsd card and I also think the connections in the microsd card slot were not the cleanest. I would eject and insert the microsd card which would sometimes fix the problem. I got rid of the supplied microsd card and purchased a new microsd card and flashed it and the system has been much more stable than the original 16gb microsd I flashed several times. New card is much faster and has more memory(not that important).
Remember, a fresh flashed microsd card will format the pcie drive on first boot.
When you get up to 50% it does slow down a lot. Change in block structure. Have you checked your debug.log file on the bitcoin directory of the pcie drive? it updates all the time as the node is running. I would open and view the changes in the log at the end of the file. That will tell you if the IBD is downloading and processing.
open an executable window -- "cd /media/pcie/Bitcoin" then "cat debug.log" will list the file and you can see and scroll through the last lines in the code which is a little faster than opening the editor.
Other optional node settings to adjust cache size would use more memory, can help the IBD and which can cause bitcoind crashes which will occur when you run out of available ram memory. bitcoind is supposed to automatically adjust to the computers available memory as the blockchain downloads. I always check the status of the log. Did you get an error or did it just drop off. IF it just drops off, then you can just restart the node and it should continue. Even with error messages, restarting the node can sometimes fix and get by the error. I keep all the keyboard HDMI windows closed and rely on remote access which uses less memory. Having the GUI opened in the browser consumed more system memory. I would keep the system monitor open so I could see the memory and cpu use. Raspberry Pi cpus are not fast on IBD.
Backup power supply. This system is very sensitive to electrical noise/surges. This can knock the raspberry pi offline and crash the bitcoind program. Any system crash can corrupt the database memory and then the simplest solution is to reformat the drive and start over. This would happen if you cant get past a node error in the debug log after multiple node restarts. Stopping the node and then restarting can also possibly fix issues, but you must make sure to wait for the node to stop and start by viewing the log. It will show you when shutdown is complete. Opening the log in an editor will sometimes alert you when the file has updated with new data as the log is written. Thus a backup pcie drive I what I go and copied the main drive with USB adapter.
)
