I have nothing in my logs either. It could have been me deleting nodes that were not doing things when I viewed them. Im not deleting anything as someone suggested ad let’s see if it works out. I have reached out to Chia on Keybase.
Try re adding me, I’ve 50 out of my normal 80 connections, I have plenty of space.
I would not go that far.
- CPU speed - you will see in Task Manager big spikes in CPU usage (I don’t see any)
- RAM size - Task Manager shows the usage, in my case it is about 50%
- Disk speed - Task Manager shows disk usage, again mine is zilch
With that said, my node is bogged down from time to time (e.g., yesterday my number of staled partials on Flex went up to 80% or so, and it was not reflected with any changes in partial processing times).
My full node has no issues like that. Although, looking at my Connection section, a lot of peers are just pulling a lot of data, while providing nothing. I hear what you are saying that some nodes are starving and are trying to catch up, but it doesn’t look that my node is starving, rather to the contrary is doing everything it can to help those starving nodes. Saying that, are we sure whether those are really starving nodes, or rather as you stated some zombie nodes that pretend to be starving?
Do not worry about frmr, at least for now it is safe, and has negligible impact.
1 Like
I don’t think so. The point of syncing up is to check what you have in your db, and pull the new stuff, not really to go over the old stuff, and fix potential issues there. It is a rather well know problem; therefore, those live backups for a given db are really helpful.
My rig is 64gb ram, ryzen 9 3950 cpu with 32 cores, 5 TiB of SSD, and 104 TiB in Enterprise SSD. I have checked every error log at any level and nothing. All I can do is wait. 
1 Like
If that doesn’t meet the minimum resources, I am selling my farm 
Could you watch some high-res YT videos (set it to 4k, if you can), and see whether it will be smooth or not (that will test your ISP connection somehow for a longer time than running a speed-test). Your rig is on very high-end of the spectrum, so no point of beating on that end. Sure, maybe you have memory or disk failing, but that can/should be tested. I still think you may have cluster damage under your blockchain db, if not, your blockchain db is just corrupted.
With that said, could you open your file explorer, right-click your c: drive (assuming boot), select “Properties”, click “Tools” tab, click on “This option will check the drive for file system errors.” It will ask you to reboot, and will take a couple of minutes.
Once you do that, type “windows memory diagnostic” in the search bar (lower left), select “Restart now and check for problems”
Those are basic (fast) checks, but having those done will remove potential questions about those components.
I did the video thing and it is smooth as can be. I also checked my disk with no errors and I just started the memory tests. Good ideas, but I fully believe this is an SW issue and I need SW help that I’m willing to pay for. I deleted the DB last Friday night and it is only at 118K I have no idea how far along chia was as the method of showing what you have and where you need to go is gone as well.
If you check the comments elsewhere, it is clear that underpowered nodes are what’s causing the current choke of the network. If you’ve tried running (and syncing) a Chia client on a 4GB Rpi, you’ll know what I’m talking about. Even during a smoothly running network.
But Larry since told us his specs, and clearly underpowered hardware is not the problem here.
@Lsherring: I still think you’re best off waiting for the current network issues to settle down. Otherwise we’re diagnosing over a known issue.
While waiting for that, I would re-install the Chia client: Uninstall in Control Panel/Programs & Features, restart Windows, then install again. In the past that has cleared some issues for me. Backup your .chia folder in between. No data will be lost, it’ll pick up where you left off after re-install. But I think maybe you already did that…
1 Like
We agree on that. The main reason I asked to do it is so you can tell everyone that those are not your issues, and you can back it up. There is just too much noise about not meeting some fuzzy criteria hysteria right now, having it done just put those things to rest.
Just a thought , but you’ve not had windows crash on you?
No, No crash. A simple upgrade I have done many times with no issues. It blew up something that Chia does not like. The memory test is running and will take several minutes as it is only 10% complete.
Couldn’t hurt to run dism and scannow just incase windows is being shitty.
1 Like
Hi, i just watched a youtube video posted today by DigitalSpacePort.
He thinks he may have a (partial) solution to some of the problems that are currently experienced with the network-issues.
As I understand his approach it is twofold: 1) upping a couple of parameters in config.yaml that rule the number of connections, so you have better chance of picking up good peers.
2) changing you’re (public) ip address, because the current one may be rejected by other peers because your system is not responding normally.
1 should be doable for most farmers i guess, 2 might be tricky, depending on your provider and type of connection.
parameters in config.yaml changed by DigitalSpacePort to:
max_inbound_farmer : 30
target_outbound_peer_count : 40
target_peer_count : 120
If those are already at those numbers, no need to change i guess, but worthwile to check…
1 Like
I am not debating that at all. I also fully agree that this is a part of the problems that Larry has. No doubt about it.
Two things that I am after is to somehow clarify what those specs for underpowered nodes are, and whether that is the only problem we have on hand.
Again, my node is fully synced, looks like not underpowered, and looks like is serving other peers with the maybe 10:1 ratio. Yet, potentially direct exchanges between my node and Flex pool are being affected. Maybe that is due to that high ratio of uploads my node is doing.
1 Like
So what happens if you do this? You may have some missed transactions.
Doing this task right now. Memory passed.
Missed transactions is one thing, broken transactions (you only have part of the record) is another thing. Once you have it in your blockchain db, whatever process that needs to get the data will get crap.
Just to copy, you really need to fully shut down chia full node. I will try to search for some live backup for sqlite, and if I find something, I will provide an update.
Ran this test with 100% success. Bringing up Chia again and it is sitting in whirly bird mode atm. Going to give it a few minutes and try again.
How it is diffrent than your computer crashing and rebooting? Database is able to sync itself from the last successful transaction.
A crash does not allow for programs and processes to close out gracefully. A reboot, allows (supposedly) for everything to shutdown gracefully and in an orderly fashion. SO your server reboots with no issues. I just tried to get Chia up and ow it wont connect to my wallet. Why did I upgrade?!