When a ChunkHeader object is constructed, it reads the chunk header information, for example the metadata position, and writes it to the log:

long c = input.readRawLong(); // chunk size
Logger.log(LogTag.JFR_SYSTEM_PARSER, LogLevel.INFO, "Chunk: chunkSize=" + c);
...

This is legacy from when JFR didin't support streaming. Today, the header is read reliably after inspecting the FILE_STATE in the refresh() method and then logged.

This means the same information is written twice, and potentially incorrect the first time. It makes sense to remove the unreliable logging to reduce noice and avoid confusion.