'GasLimit' field cannot be greater than a block's gas limit! #9103
Comments
|
Additional note: This deadline (2) was not compacted but deadlines 0 and 1 had been compacted last week. |
|
Did you set the MaxPartitionsPerPoStMessage And MaxPartitionsPerRecoveryMessage? |
|
I would recommend lower it, especially the recovery one first! |
|
I only have one partition (first four deadlines) |
|
BTW - I found it in lotus-miner config default. |
|
It is important to note that I have been running v1.16.1 without problem for the last few weeks and this problem only showed itself two days ago on July 29. (So today is my second day of faulting on deadline 2.) Frankly this makes no sense to me. I only have one partition and changing from default (0 unlimited) to '1' should make no difference. I did now add the [Proving] block and reset the value to '1'. This is going to keep costing me penalties until this bug is fixed.
|
Reiers
added
area/proving
|
Cost so far
|
|
New .lotusminer/config.toml settings
|
|
First of all, this issue is above my understanding of how the GasLimit is calculated, so these are just some observations for further context:
Do you have the epochs before and after the Changing from 0 to 1 should not have any impact at all, since you only have 1 partition. Changing this value only makes sense when you have more then 2-3 partitions in a deadline. Only thing I can think of that maybe could be beneficial before someone more knowledge about GasLimits takes a look at it is to enable LOTUS_VM_ENABLE_TRACING=1 to get traces in case it happens again. |
I have since restarted the lotus-miner with the new settings so no older logs now. After the wdpost fails again in a few hours I can collect more log info. I also added the envvar "LOTUS_VM_ENABLE_TRACING=1" to the miner/lotus node. |
|
Additional context July 28 loss of all power due to Internet outage on all four deadlines (chain not in sync unable to PoST any messages) |
|
Start of wdpost for July 31 (both deadline 0 and deadline 2 (DeclareFaultsRecovered for 2125 sectors) 2022-07-31T14:53:36.053Z WARN storageminer storage/wdpost_run.go:237 Checked sectors {"checked": 1443, "good": 1443}
|
|
July 31 Deadline 2 wdpost 2022-07-31T15:48:37.957Z WARN storageminer storage/wdpost_run.go:237 Checked sectors {"checked": 2129, "good": 2129}
2022-07-31T15:58:30.893Z ERROR storageminer storage/wdpost_run.go:181 submit window post failed: pushing message to mpool:
2022-07-31T15:58:32.272Z ERROR storageminer storage/wdpost_run.go:181 submit window post failed: pushing message to mpool:
And continues to repeat.... |
|
|
|
Lotus log - select messages 2022-07-31T15:58:19.414Z WARN fullnode full/mpool.go:171 Push from ID address (f03197), adjusting to f3q3crofnq3sxmbyqwxsx7qcmjuykr4uk7jgcrxlrdrk7ovjkmlbhgqhkcomeeqcum5l6h4pugxigrbolz5fkq 2022-07-31T15:58:30.892Z WARN fullnode full/mpool.go:171 Push from ID address (f03197), adjusting to f3q3crofnq3sxmbyqwxsx7qcmjuykr4uk7jgcrxlrdrk7ovjkmlbhgqhkcomeeqcum5l6h4pugxigrbolz5fkq |
|
Cost so far -102.7519 FIL x '$10.30 = < $ 1,058.34 >
|
|
A temporary remedy: maybe you can reduce the number of sectors in |
|
James, how would I do that? |
|
@stuberman we are looking into this rn and hopefully give you some tools, soon |
|
With #9109 it looks like your sector number allocations are pretty good, which invalidates one of the theories as to why gas use is so high here The second partition has ~400 runs in it's bitfields, which isn't much different from other miners, so there may be something else going on here |
Checklist
Latest release, or the most recent RC(release canadiate) for the upcoming release or the dev branch(master), or have an issue updating to any of these.Lotus component
Lotus Version
Describe the Bug
Unable to post one of 4 deadlines two days in a row:
See closed issue 8997
lotus-miner proving infolotus-miner proving deadlinesMiner: f01278
Logging Information
Repo Steps
See closed issue -this is not resolved in v1.16.1
#8997
The text was updated successfully, but these errors were encountered: