home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!mcsun!julienas!jussieu!card
- From: card@masi.ibp.fr (Remy CARD)
- Subject: Re: Inode/Zone count errors with ext fs
- Message-ID: <1992Nov17.161106.18763@jussieu.fr>
- Sender: news@jussieu.fr (Le Facteur)
- Nntp-Posting-Host: ares.ibp.fr
- Organization: Laboratoire MASI - Universite Pierre et Marie Curie - Paris - France
- References: <1992Nov14.184337.23031@jussieu.fr> <1992Nov14.225450.23196@prime.mdata.fi>
- Date: Tue, 17 Nov 1992 16:11:06 GMT
- Lines: 27
-
- In article <1992Nov14.225450.23196@prime.mdata.fi> kennu@mits.mdata.fi (Kenneth Falck) writes:
- >
- >I wonder, what would happen if you first sync'ed (three times :),
- >then ran efsck, and right after that switched the power off? Is
- >there any other data kept in the kernel buffers after a sync except
- >the superblock?
-
- If you sync, all buffers are written back to the disk and
- the memory copy of the super block is also updated on the file system.
- When you run efsck, it can correct the super block on the disk. If
- you switch off the power just after running efsck, the memory copy of
- the super block is not written to the disk so the changes made by
- efsck are not overwritten and you win.
-
- Note that if a process allocates or deallocates a block
- or an inode just after you ran efsck and before you switch off the
- power, you lose because the memory copy of the super block has been
- modified.
-
- >
- >--
- >kennu@mits.mdata.fi
-
- --
-
- Remy Card
- card@masi.ibp.fr
-