[ERR] BTCD: Unable to start server on [:8333]: EOF

[abitrolly]

While experimenting with different revisions of btcd to catch the cause of #1024 I faced with this error

2018-04-08 18:18:19.704 [TRC] BCDB: Scan found latest block file #150 with length 150086231
2018-04-08 18:18:19.705 [INF] BTCD: Block database loaded
2018-04-08 18:18:19.712 [TRC] CHAN: Serialized chain state: d2b49fbdefd1492deed353a70f197330da1cac980aabae030000000000000000aa7f0600d8bef608000000000b000000238e9e08e1475e7f176b73
2018-04-08 18:18:19.712 [INF] CHAN: Loading block index...
2018-04-08 18:18:23.141 [ERR] BTCD: Unable to start server on [:8333]: EOF
2018-04-08 18:18:23.141 [INF] BTCD: Gracefully shutting down the database...
2018-04-08 18:18:23.142 [INF] BTCD: Shutdown complete

From other tickets it looks like jumping to an older version may somehow corrupt DB. If it is the case, could the error message be improved?

[stevenroose]

@abitrolly For one, it is true that jumping back to older versions can make the DB seem corrupt. Those occasions are usually mentioned in the PR introducing them, but I agree that using explicit versioning to indicate such would be a good idea.

As for the EOF: it looks like another program (another btcd/core daemon?) is already using the port.

You can change the listen port with --listen=0.0.0.0:9333, f.e.

[abitrolly]

@stevenroose I double checked that nothing was running on port 8333 at that time, that's why I concluded that the error message is misleading. True (unable to start), but missing the right reason.

[meyer9]

EOF during startup like that generally means it had trouble reading the block index.