Head of blockchain fall back to 0 after unsafe shutdown, how to reset it to the latest?

Question

System information

Geth version: 1.8.27 OS & Version: Linux Docker version: 18.09.0

Background

gcmode is full, syncmode is fast. Single node in private network. datadir mount to docker volume.

Problem

I restart the docker so that geth quit unsafely

sudo service docker restart

When I restart the geth, the latest block become 0.

eth.blockNumber
> 0

it was 140000, so 140000 minded blocks are missing. However, I still can get data of old blocks

eth.getBlock(140000);
> {data of this block}

So it looks like blocks still exist but the head of chain reset to 0.

Questions

1 Does full gcmode cause this issue?

2 Is there a way to reset the head? I tried

debug.setHead("0x222E0");

but nothing happen.

Answer 1

I had the same problem, I tried everything. The only way worked:

1. export all block data by run golang code

chainClient, err := ethclient.Dial("./db/geth.ipc")
if err != nil {
    log.Fatal(err)
}
fh, err := os.OpenFile("./aaa", os.O_CREATE|os.O_WRONLY|os.O_TRUNC, os.ModePerm)
if err != nil {
    log.Fatal(err)
}
defer fh.Close()

var writer io.Writer = fh
for nr := 0; nr <= ___ 140000____; nr++ {
    blockNumber := big.NewInt(int64(nr))
    // blockNumber.Int64(nr)
    block, err := chainClient.BlockByNumber(context.Background(), blockNumber)
    if err != nil {
        log.Fatal(err)
    }
    if block == nil {
        log.Fatalf("export failed on #%d: not found", nr)
    }
    if err := block.EncodeRLP(writer); err != nil {
        log.Fatal(err)
    }
}

2. geth init
3. geth import aaa

Then everything is right. eth.blockNumber is no longer zero.