I have now set up two DIY Start9 servers and both seem to have this RPC Timeout problem… A lot… I have the entire blockchain synced. When the error isn’t showing it reports that all is synced, etc. But every time I look at Services I see the yellow exclaimation marks on Knots, Electrs, LND, Mempool. Go into any of those services and all report the same thing; “Timed out, retrying soon.”
What is with these Sync errors? Is there some setting I can adjust to make it work more reliably?
My machine has 16Gb of RAM, a 4 Tb hard disk. Intel i3. Ethernet cable (not wifi) connected via Starlink.
Do other node servers suffer this problem? Umbral? Others?
This is an error that points to lack of resources. The RAM you have looks fine. A spinning magnetic hard disk does not have the read/write speeds to run Bitcoin, so you’d need to replace that with an SSD, and an “i3” doesn’t really say much, since those span about 15-20 years.
But certainly a resource bottleneck somewhere.
A spinning magnetic hard disk does not have the read/write speeds to run Bitcoin,
That really blows me away considering a block is only written once every 10 minutes or so. And it doesn’t bode well for the global south ever being able to a run a node if you have to have a SS HD.
I’ll try closing down all other services except Knots and Electrs in the meantime while I drime up an SSD.
Thanks for the reply.
OK, so you were right, it is a resource issue. I can run Knots and Electrs no problem with no PC Timeouts… But as soon as I run Mempool, I get the almost permanent RPC failures. Running RoboSats does not cause the problem.
It’s on a i3 “Kabylake”… whatever that is.
mempool.space is a very heavy service, as it builds additional indexes of both blocks and the mempool, putting constant strain on cpu, memory and the storage device.
You’ll definitely need a good CPU and fast SSD/NVME drive to be able to run mempool.space