Failed to truncate extra state histories err="EOF

[i11m20n]

System information

Geth version: 1.14.7-stable
OS & Version: CentOS Linux release 7
Commit hash : aa55f5e

Expected behaviour

Start syncing blocks

Actual behaviour

Exit after CRIT

Steps to reproduce the behaviour

After power failure, restarting geth encountered this problem

Backtrace

INFO [08-21|10:05:19.338] Logging configured                       rotate=false format=terminal location=/home/fengfeng/.ethereum/geth.log
INFO [08-21|10:05:19.354] Starting Geth on Ethereum mainnet...
INFO [08-21|10:05:19.354] Bumping default cache on mainnet         provided=1024 updated=4096
INFO [08-21|10:05:19.386] Maximum peer count                       ETH=100 total=100
INFO [08-21|10:05:19.389] Smartcard socket not found, disabling    err="stat /run/pcscd/pcscd.comm: no such file or directory"
INFO [08-21|10:05:19.390] Using pebble as db engine
INFO [08-21|10:05:19.395] Set global gas cap                       cap=50,000,000
INFO [08-21|10:05:19.395] Initializing the KZG library             backend=gokzg
INFO [08-21|10:05:19.465] Allocated trie memory caches             clean=614.00MiB dirty=1024.00MiB
INFO [08-21|10:05:19.525] Using pebble as the backing database
INFO [08-21|10:05:19.525] Allocated cache and file handles         database=/home/fengfeng/.ethereum/geth/chaindata cache=2.00GiB handles=2048
INFO [08-21|10:05:20.780] Opened ancient database                  database=/home/fengfeng/.ethereum/geth/chaindata/ancient/chain readonly=false
INFO [08-21|10:05:20.782] State scheme set by user                 scheme=path
INFO [08-21|10:05:20.797] Initialising Ethereum protocol           network=1 dbversion=8
WARN [08-21|10:05:20.797] Sanitizing invalid node buffer size      provided=1024.00MiB updated=256.00MiB
INFO [08-21|10:05:21.863] Failed to load journal, discard it       err="unmatched journal want d16c808bc3a98178feb934e29fd7c92eca38634a1e802357d9a9971f3048d2eb got cfdf68c515b377991f0508611f849eccc2075ed9c995e67a69ce1b2f19b1324b"
ERROR[08-21|10:05:21.865] Corrupted index file detected            lastOffset=0 indexes=11,756,027
ERROR[08-21|10:05:21.865] Corrupted index file detected            lastOffset=0 indexes=113,247
INFO [08-21|10:05:21.866] Opened ancient database                  database=/home/fengfeng/.ethereum/geth/chaindata/ancient/state readonly=false
CRIT [08-21|10:05:21.892] Failed to truncate extra state histories err="EOF, fileid: 0, start: 858186321, length: 3436780975"

[fjl]

The issue is a corrupted index file. At the moment, Geth cannot recover from this database corruption and you have to resync. However, we have a pending PR to improve handling of such errors: #29792

You can try building geth from the branch of #29792, and start it one time. Theoretically, it should detect this corruption and revert the database to the last good state. Worst case, if it doesn't work, you have to resync.