hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hairong Kuang" <hair...@yahoo-inc.com>
Subject fsck with the move option
Date Mon, 01 May 2006 21:53:24 GMT
After I read the fsck code, I realize that if we run fsck with the move
option, the tool actually copies the corrupted files to "/lost+found" block
by block. I feel that it is very expensive especially with large corrupted
file system. Shall we simply do a rename, I mean, moving corrupted files to
"/lost+found"? Rename is a namespace mutation so it is much more efficent.
 
Hairong

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message