Someone keeps dusting the chia network -

Some explanation for current slow transaction processing, sync issues, logs overflowing with warnings.

1 Like

Would this explain why a transfer that I made 10 hours ago still hasn’t shown up in my cold wallet ?

Could be. I did a transfer that took over 4 hours, just came in 1/2 hour ago.

Would it also explain stale partials I’ve had the last couple of days when I mostly don’t get any?

Also noticed the fan on my farming laptop kicking in a lot in the last couple of days.

1 Like

Same here, both!
We are not the only ones… Netspace is going down, on my (Space) pool this also is visible. All estimates, influenced by this issue, are off it seems.

80% of the peers are not full nodes. They do not reach Peak Height.

Space pool seems to have delayed paying out today as well (for me anyway) .

I temporarily turned off wallet sync. Only node sync and farm are open.

For those that have problems (I did), I would suggest that you do two things. First is reduce the log level to ERROR. In my case, that was fully choking one CPU core that was used by start_full_node process. Once chia got restarted, things started moving faster.

The second is to reduce peer count down to 10. Part of the problems on the node is synchronization between those peers and db access. When things go bad, chia cannot handle that synchronization properly and starts dropping data that just received. By reducing the peer count, the db pressure is relieved a bit.

With those two changes, my node runs fine right now, no single core is choking, I (almost) don’t have stales anymore.

1 Like

If you do a manual xfr, always add a small (1 mojo or so) fee for that transaction. This boosts priority of such transaction. Basically, what I do is to first make a transaction for 10 mojos or so, with 1 mojo fee, just to see whether it goes through, and then follow that with the full amount with the same fee.

In order to run something like dust storm, they either need to run plenty of free xfrs, or have some chunk of money to spend (but that could be exhausted rather quickly). So, putting some fee on your manual xfrs basically puts your transaction at the head of the processing queue.

By the way, this attack (or rather this wave) started about 10 hours ago, so yes you got caught in this mess.

1 Like

Aren’t all nodes full nodes? They can be behind, but they are still full nodes. Also, how did you get that 80% number?

One thing is that if your node is affected, it basically will stop reading status of other nodes. Therefore, the fact that you see 80% nodes behind may be that your node is not updating those height counters properly. That was what I had when my node was struggling.

Thanks, yes I made the transfer with a fee of 1 mojo. Checking the debug log it shows a wallet warning but the transfer was confirmed at the next height. I suspect it will take some time to shake out. Bad timing to win a block on my part :slight_smile:

number of peers is 8. I limited it. If the peer doesn’t have a Peak Height value, I delete it.

target_peer_count: 8 in config.yaml

This really good for me. But you should always watch the peers. So slow ones don’t affect you.

Again, peers reflect the state of the network. If 20% of nodes are screwed up, that what you will end up seeing with your peers. If you start knocking off those peers, you just get a fulltime whack-a-mole job. Just let it sort itself out, give it some time to settle down. Those nodes that are really struggling will eventually drop off / time out, and new nodes will come.

Also, there is a difference between those nodes with none or 0 height, and those that are behind your node. The first group implies that your node has issues, the second group is a normal thing, so you should not act on that group. You only need one/two good nodes to sync. If you let few other nodes with a lower height hang on, you are potentially helping them to get synced.

Yeah, there is a third group as well. Those nodes that are more behind, but their counters are not updating anymore. Those peers are handled in the same way as those with none/0 height - will time out at some point, and be replaced. I would not worry about those as well.

Does the concluding paragraph of the piece lead to an assumption that hobby farmers, on older PCs and the Pi crew, will be muscled out from Chia space?

" If it is just low powered nodes, then they need to go. Permanently. "

1 Like

I can’t follow the chain height for many days now. I even put my database on an SSD with 5950X CPU, 1Gb/s fiber internet also, and I still can’t sync.

Chia devs should fix this as soon as possible. The current transaction time is very slow for the users. For the miner, it’s even worse. Please do your job, you got a lot of pre-farm for that didn’t you!!!

1 Like

Could you take a screenshot of your Connections section from the Full Node panel and post it here?

That is just a BS pushed to have you not comment on shit code chia has. If someone mentions “low powered” nodes, but is not specific whether that is RAM or CPU or number of cores or media holding blockchain db, … that is just an utter BS. The code is so deficient that is not worth to comment on it.

On the bright side, Flex is going to have a full farmer potentially soon. There are two things about that. The bad part is that their current farmer implementation is not really as chia would like to see. Some people see it as a potential problem for the network. On the other hand, if they can easily handle a full node on RPi3 and those other “low powered” devices that would call the bluff, and force chia to get their act together.

Another good thing about it is that once that Flex node will be released, people with those “low powered” nodes will have to make a decision - upgrade their H/W (not knowing what really works) or just switch to Flex. If folks will go with the second option (I will), other pools will see their net space dropping, as such will start to scramble for a solution.

I started to have sync drops Friday morning (Dec 17) which came and went up to a real kick in the teeth around noon est today that dropped me down about 1,000 from the peak. Almost caught up now (within 200) . I pool at the moment that gets me .01 or 0.02 chia a week.

I am a hobby farmer with 20TB of space, 14TB full, that is being methodically populated at 3.5 or so plots a day. Just using an old i3 4 core with 14gb of memory. I plan to buy 6 or so TB a month up maybe 60 to 100TB

Where would one keep up on developments that would give a heads up on new required hardware?

Look at some other today’s threads, and hopefully you can get your node running better (less stales).

Actually, here is the official page with reference H/W:

That’s the thing. When one says what is considered “low powered” problem that implies that the problem is understood, and either it can be addressed, or is rather critical (for instance a low-speed SD cards on RPi are critical) On the other hand, when people just talk about “low powered” boxes in general, it implies they have no clue what is actually working or not, as such it is just a cover for not understanding the problem.

So, I would rather not worry about the issues that are out there, as those will be sorted out one way or another.

Basically, there is more than 300k nodes / farmers out there and saying that the company has no resources to get the shit fixed is just a nonsense.