2

I recently moved the first 2000 files of my blockchain from the SSD to a HDD. The HDD gets mounted in /mnt/ext.

A symlink per file is then put into blocks that points to the remote files. So my blocks folder is still on SSD, but the files are on HDD.

I tried a bitcoind -reindex-chainstate to be sure everything still works but now it is giving an error:

2020-11-26T12:52:24Z ERROR: ReadBlockFromDisk(CBlock&, CBlockIndex*): GetHash() doesn't match index for CBlockIndex(pprev=0x44288c0, nHeight=255, merkle=4309bfeed77a70f309da08bcf8948906b9cc26120c0b0ef86e0ac67284bbd79e, hashBlock=00000000d0a75c861fabf9ff7b92022f60e4afeed9331fe5aa073d8e4706fe3c) at FlatFilePos(nFile=0, nPos=58808)
2020-11-26T12:52:24Z *** Failed to read block

Now I have several questions. I think nFile=0 means that the block is in blk00000.dat and rev00000.dat. Am I right?

So I copied this file again from a backup of the blockchain. This is from a second node. But I would suspect the files are the same, right?

But I get the exact same error at the same position - even with a new -reindex-chainstate.

I can see in the log that all blocks up until number 254 get indexed but each time it fails on block with height 255.

I think those old files contain always the same blocks. How can it be 254 blocks are correct but it fails on the next one?

Could this simply be a problem of the mount?


EDIT: I am currently -reindexing. In this case the files I already have linked to from my blocks folder get used like this:

Reindexing block file blk00000.dat...
Loaded 119964 blocks from external file in 133291ms

Apparently, the first file contains 119,964 blocks. So when I have an error with height=255 it should definitely be in the first file.

3

So I copied this file again from a backup of the blockchain. This is from a second node. But I would suspect the files are the same, right?

No. Blocks are downloaded out of order, but are written to disk in the order that they are downloaded. So the block index (which states where each block is located) will be different for every node because every node will likely have the blocks stored in a slightly different order.

Furthermore, if your node has received stale blocks, those stale blocks are also stored on disk and have an entry in the block index. Not all nodes will have the same stale blocks, and they may not be stored in the same order. Newly synced nodes will not have stale blocks from before they were synced, and often times stale blocks become stale due to propagation delays, so a node may not even receive a stale block to store.

You can't just copy individual block files from one node to another. You need to copy the entire blockchain and the block index. Otherwise you will probably corrupt the whole thing and need to redownload the blockchain.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.