Cannot remove sectors stuck in Committing #6522
-
After 0.8.0 upgrade, many miners reported that their sectors are stuck in Committing status(for days), and when they try to remove it using So I tried to reproduce the issue, and the remove does work after a while when the following shows in the log 0-09-28T23:55:43.296-0400 ERROR evtsm [email protected]/machine.go:83 Executing event planner failed: running planner for state Removing failed:
github.com/filecoin-project/lotus/extern/storage-sealing.(*Sealing).plan
/Users/jennijuju/filecoin/software/lotus/extern/storage-sealing/fsm.go:205
- planner for state Removing received unexpected event sealing.SectorCommitted ({User:{Proof:[175 221 6 162 182 25 128 3 115 35 185 16 162 68 101 206 189 78 215 111 209 138 170 171 252 214 190 160 212 57 16 212 64 41 9 91 73 60 200 177 246 89 34 3 209 227 66 143 181 65 35 219 168 142 68 9 80 18 165 208 49 58 200 231 236 117 134 240 60 69 239 255 4 71 132 72 162 128 214 25 234 66 220 86 76 86 96 171 34 246 180 63 23 199 198 19 21 97 124 132 72 133 98 171 107 4 116 109 37 105 208 135 68 220 96 197 70 194 212 52 110 123 216 43 101 60 233 209 166 40 177 71 188 249 232 245 31 124 129 130 139 47 102 224 179 47 239 72 107 157 64 231 145 241 206 20 150 68 105 110 36 178 143 198 245 228 219 148 152 231 235 100 213 129 241 93 93 91 17 97 135 112 86 153 218 172 227 146 181 74 214 134]}}):
github.com/filecoin-project/lotus/extern/storage-sealing.planOne.func1 And the status of the sector is moved to However, if we check the log of the sector, it seems like it is removed after Committing was successful and this is may be why the miner cannot remove it(since they arent able to finish committing)
So the question/issue is:
|
Beta Was this translation helpful? Give feedback.
Replies: 8 comments
-
dump from slack
@magik6k can a proof generated by the seed before upgrade still be committed after the upgrade epoch for this case? |
Beta Was this translation helpful? Give feedback.
-
Here are my issues about faulty sectors: e.g I have 2 sectors (40 & 41)
I am trying to to remove them with 2 methods:
Method2:
And the sector status is following:
|
Beta Was this translation helpful? Give feedback.
-
turns out sectors in commitwait cannot be removed either(Note: files for this sector are removed manually, too). seems like there is a pattern that sectors are stuck in SectorRestart. |
Beta Was this translation helpful? Give feedback.
-
It'd be great if we can get this looked at soon, since this is one of the top5 pain point for miners rn. |
Beta Was this translation helpful? Give feedback.
-
ID State OnChain Active Expiration Deals DealWeight
lotus-miner sectors update-state --really-do-it 20 Removing Doesnot know if it is related. |
Beta Was this translation helpful? Give feedback.
-
I did find a work around for this issue if anyone needs assistance in this. I then shut down my miner and made a back up. After that I rebooted miner and it then found the bad files and removed the sector. 2021-02-11T15:45:36.922-0600 INFO stores stores/remote.go:334 Delete http://10.0.120.250:3456/remote/sealed/s-t047419-630
After this I then saw a state of removing and it was gone from the scheduler. |
Beta Was this translation helpful? Give feedback.
-
This solution presented above has worked for me after experiencing this exact issue |
Beta Was this translation helpful? Give feedback.
-
Thats a fun workaround, but shouldn't we just be able to terminate sectors properly? |
Beta Was this translation helpful? Give feedback.
I did find a work around for this issue if anyone needs assistance in this.
After some research I found that the miner was running C1 on these sectors and it was then getting stuck looking for the data to move forward to the next step.
Mine was for sector 630
I then shut down my miner and made a back up.
I created a dummy file in the sealed/ folder of my sealing repo called s-t047419-630 (touch /sealingrepo/sealed/s-t047419-630)
I also created a empty dir in the /cache folder of my sealing repo called s-t047419-630 (mkdir /sealingrepo/cache/s-t047419-630)
After that I rebooted miner and it then found the bad files and removed the sector.
Here is the log after miner rebooted and finally re…