Yes this is exactly what happened with me and was repeatable after starting again from scratch. At that point after a couple tries, I gave up and simply copied over a data directory to get it running.
I also tried syncing on Amazon a1 instances Amazon EC2 A1 Instances. The a1.medium (1 CPU, 2GB RAM) didn’t work (albeit I didn’t enable any swap) but the a1.large (2 CPU, 4GB of RAM) worked great so I think the Pine64 should work really well and the Odroid C2 is right on the edge.
I have a plan, its a plan so cunning that if it had a tail it’d be a fox…
Started up again with a much smaller dbcache size (8) - the default is 450 - with a little luck that’ll free up enough resources to get through the sync.
This happens because updating the metrics screen requires locking on cs_main, and if the node is busy processing a particularly-large block, it will end up holding that lock until the block is finished, meaning that the UI will hang until its done.
Mine crapped out around block 115000, the Pi3B+ power supply doesn’t have enough grunt to run my external hard drive, which is essential for a swap file. When swap activity increased from digesting larger blocks there wasn’t enough power left to run the Pi.
I had it running for a while with swap enabled on the SD card (epicly bad idea!), it worked but was very slow. I’m going to hack the USB cable so the drive can have its own power supply & try again.
(Edit: Overclocked the SD card (heh!) & stole some RAM back from the GPU…trying again)
Not too bad so far, average 721mb RAM use, 26k blocks. But it’s going to take me a while no matter what because I live in a rural area with embarrassingly slow internet speeds.
It has 4GB of RAM, (3.79 free with LXDE desktop) so hopefully it won’t choke on the big blocks.
I came home to find that the node had stopped at 49K blocks and had no connections. I’m not sure what happened but the entire board lost internet sometime last night or early this morning.
So I re-booted the router and the board and now everything is back to normal… so far…
I could just copy the blockchain from one of my other nodes, but I was hoping to find out what a new user would experience if doing this for the first time.