Re: in case of disk failure,why not recover from middle position of the log file

Hi Zhang, Yes, when doing log recovery, we'll start from the checkpoint for the partition in `recovery-point-offset-checkpoint` file, which will be updated at runtime. Is that what you expected? Check LogManager#loadLogs method for more implementation detail. Thanks. Luke On Thu, Oct 13, 2022 a

in case of disk failure,why not recover from middle position of the log file

can we write the byte position of .log 、.index file to a position-checkpoint file, and recover from the position to rebuild the index files?