Open alnuweiri opened 2 years ago
Also it is worth mentioning that when this happens I will get twice this message in the log "XX plots were eligible for farming..." where the second time it's always a much faster lookup time for the plots:
2022-04-25T00:20:57.985 full_node chia.full_node.full_node: INFO Finished signage point 60/64: CC: 5805e47d0e302987d1c006034ae9fa9534dd536e1831c76a1b82e119cc41d193 RC: fddbe670cccd7914646a221495f19e911099716fb74eeb6a0c6073631c4fd281
2022-04-25T00:20:58.117 full_node chia.full_node.full_node: INFO Updated peak to height 1884395, weight 4013880992, hh 83b9c07e24c0f46d025379a48596e46d51119b46e26767c94d794838383def99, forked at 1884394, rh: 969fbfd2cbf18224ebf0945e94ae928a94c82b27e698bcb15211baecc8635796, total iters: 7434506884419, overflow: False, deficit: 0, difficulty: 2352, sub slot iters: 135266304, Generator size: 73067, Generator ref list size: 0
2022-04-25T00:20:58.124 full_node chia.full_node.full_node: INFO Finished signage point 60/64: CC: 5805e47d0e302987d1c006034ae9fa9534dd536e1831c76a1b82e119cc41d193 RC: fddbe670cccd7914646a221495f19e911099716fb74eeb6a0c6073631c4fd281
2022-04-25T00:20:58.137 full_node chia.full_node.mempool_manager: INFO Size of mempool: 200 spends, cost: 107327646146 minimum fee rate (in FPC) to get in for 5M cost tx: 0
2022-04-25T00:20:58.195 harvester chia.harvester.harvester: INFO 6 plots were eligible for farming 25002045c2... Found 0 proofs. Time: 0.19908 s. Total 2533 plots
2022-04-25T00:20:58.203 harvester chia.harvester.harvester: INFO 6 plots were eligible for farming 25002045c2... Found 0 proofs. Time: 0.04967 s. Total 2533 plots
After investigating more on the issue, I found that the problem is caused by receiving the signage point twice for many of the signage points. This problem is affecting many people, for example, there is one post on Reddit where a person found two proof in less than a minute but received only one reward, after checking his log file I found actually that reason is that he got the signage point twice and the harvester checked twice for the proof. Also, issue #10827 is caused by the same problem. Moreover, I guess some of the people who recently reported a longer lookup time for the plots are getting this problem because the harvester is checking for proof twice in less than 1 second when they get the signage point twice.
This problem still exists even after updating to Chia 1.6.2 on Windows!
This issue has not been updated in 14 days and is now flagged as stale. If this issue is still affecting you and in need of further review, please comment on it with an update to keep it from auto closing in 7 days.
Still exists
Could you put your log into DEBUG mode and provide us with the full log?
I will do that in few days!
This issue has not been updated in 14 days and is now flagged as stale. If this issue is still affecting you and in need of further review, please comment on it with an update to keep it from auto closing in 7 days.
Here is the full log file with Debug level!
To find the problem please search for "Don't have signage point" you will find a few hits, check the first one, and you will find before it "Finished signage point 5/64"
2023-03-05T17:21:01.091 full_node chia.full_node.full_node: INFO Finished signage point 5/64: CC: ef7314e3502d06a2989d40a849a69f642e4cb2d90f83447e63c1d24d3231732c RC: 5f85479fa9694c5d58fa8b02cda500a10a823bdff2a87bd41ba3cf445b041f7c
and some lines after it again "Finished signage point 5/64"
2023-03-05T17:21:01.816 full_node chia.full_node.full_node: INFO Finished signage point 5/64: CC: ef7314e3502d06a2989d40a849a69f642e4cb2d90f83447e63c1d24d3231732c RC: 5f85479fa9694c5d58fa8b02cda500a10a823bdff2a87bd41ba3cf445b041f7c
You can see in the logs this
2023-03-05T17:21:01.459 full_node chia.full_node.full_node_store: DEBUG Reverting 5 13790390190080 13790389232797
and
2023-03-05T17:21:01.930 farmer chia.farmer.farmer : DEBUG Duplicate signage point 5
Here is another different example, I have uploaded a part of the log file as .log and the full file is uploaded as .zip because of the file size limitation on GitHub.
Here you will find this case: "Finished signage point 24/64" and then "Finished signage point 25/64" and after a few seconds you will see again "Finished signage point 24/64" and "Finished signage point 25/64"
2023-03-05T17:03:44.599 full_node chia.full_node.full_node: INFO Finished signage point 24/64: CC: ab9275652e0a780a580cf9507933171f6c128dc1e0ba023e30ca93e20e16d3da RC: 3b7fc13240ab1fe72601e6508af9ecfef9b00dbc4b98faa147a4a1b9481f192a
2023-03-05T17:03:53.532 full_node chia.full_node.full_node: INFO Finished signage point 25/64: CC: 8baa84219dd71a59272e3a1921cf52283e03da956a359abcf6b01308fcc25e6a RC: b9268b3c6d443173269d166790f91a145453a41261425b78b2d8d53c8f3a807f
And again after few seconds
2023-03-05T17:04:05.812 full_node chia.full_node.full_node: INFO Finished signage point 24/64: CC: ab9275652e0a780a580cf9507933171f6c128dc1e0ba023e30ca93e20e16d3da RC: 8dfdab7ebd326be64f78b05e2b563f116751bb0427a829b5eee4875591c336fc
2023-03-05T17:04:14.568 full_node chia.full_node.full_node: INFO Finished signage point 25/64: CC: 8baa84219dd71a59272e3a1921cf52283e03da956a359abcf6b01308fcc25e6a RC: 44b5a655431d64dc113993139c674fad9e0440494444a388d072e1f387f7c318
However, in this case, you can notice that "RC:" between the first "signage point 24" and the second "signage point 24" is different, also it's different between the first "signage point 25" and the second one.
Also, you can see in the logs these lines
2023-03-05T17:03:57.075 full_node chia.full_node.full_node_store: DEBUG Reverting 24 13790140170240 13790138150366 2023-03-05T17:03:57.076 full_node chia.full_node.full_node_store: DEBUG Reverting 25 13790142464000 13790138150366
@MumfMeisterT
Thank you @alnuweiri - Assigning this ticket out to look through the logs and get back to you.
Thanks for the logs, I took a look and it seems like somewhere in the network, signage points and a block got out of order. The block arrived after the 24/25 signage points instead of before:
2023-03-05T17:03:57.070 full_node chia.full_node.full_node: INFO Updated peak to height 3336842, weight 6824429736, hh f95b226c287a9f6a6616d58216b4fba26acb14566898cc3ca9dbb38a71db9ce5, forked at 3336841, rh: 804a5694ad517b824323d60458fd67db29d9cb1b8ef55f229044647107545765, total iters: 13790138150366, overflow: False, deficit: 0, difficulty: 1976, sub slot iters: 146800640, Generator size: No tx, Generator ref list size: No tx
This changed the reward chain, so signage points 24 and 25 were invalid, thats why they were reverted back to iter 13790138150366.
Then new signage points 24 and 25 with the altered reward chain arrived.
Now, there is an issue here. I don't think the farmer should need to do a plot lookup again (which I think is your complaint). All that has changed is the reward chain. So maybe we can cache the lookup to avoid that overhead in case this happens. Leaving this issue open for that improvement.
What happened?
I am seeing in the log a lot of "INFO Don't have signage point {'challenge_hash': ......." although if you look at the logs it seems that I already have the signage point. In the example I attached below you can see the following:
INFO Finished signage point 22/64:
So I assume that I have the signage point 22 but after that, I see this log
I get this message a few times, then I get this log again:
INFO Finished signage point 22/64:
I am wondering what is happening here, and why I am getting the signage point two times. I noticed this behavior since I updated Chia to 1.3.x
I am getting a lot of this example although I have a high-end PC with a fast NVMe disk.
Version
1.3.4
What platform are you using?
Windows
What ui mode are you using?
GUI
Relevant log output