I've experienced a strange problem on my 2.0.31-pre2 system.
After erasing two huge files (630MB) on an otherwise empty Ext2 formatted
SCSI disk, I was unable to unmount it -- device or resource busy.
I thought it was, as usual, a VC with a shell having a working directory
inside this mounted partition - but no, no such thing. A process referencing
/dev/sdc2 (the "offending" partition), but neither, lsof reported nothing
related to sdc. Looked at ps aux to see if a process was hung in the D state,
once again nothing. Switched to single user, but still unable to unmount this
/dev/sdc2 ... weird. A shame, I forgot to do a df to see the number of free
blocks remaining on /dev/sdc2 for latter references.
I remember a thread some weeks ago about a problem of deletion of huge files
on ext2 not freeing all blocks -- Anyone having a clue ?
System is bleeding edge Debian hamm (developement, future 2.0), running
libc5.4.23 and glibc2.0.3 (not so much compiled against it, except ProFTPD,
really neat if you want to hose your wtmp :).
I have three patches applied in addition of 2.0.31-pre-2:
- ncrBsd 1.18i
- Solar's stack not-exec patch
- Stephan latest IO-handling patch for Ext2
Any idea?
TIA, Philippe.
-- Philippe Strauss <philou@lili.urbanet.ch> Urbanet SA, Vallombreuse 51, 1004 Lausanne -- 021 / 625.28.14Homepage & PGP key: http://lili.urbanet.ch
Never insult an alligator until you've crossed the river.
--