. and .. get blown away when directory size > 65536 UPORT286

Eric Beser ebeser at ajpo.sei.cmu.edu
Tue Jun 28 23:28:48 AEST 1988


In article <2104 at sugar.UUCP>, karl at sugar.UUCP (Karl Lehenbauer) writes:
> I'm not exactly sure that this is what happened, but it appears that when
> a directory reaches 65536 bytes in size under Microport System V/AT it
> wraps around and the next two files created cause . and .. to get blown
> away, trashing the directory.
> 
I have had this problem when usenet sends large amounts of files. I have
found  a multitude of problems caused by this. I noticed that other
directories in the inode chain get blown away too.

> I used fsdb to change the size of the directory to zero then let fsck trash
> all the files.  I think it would be possible to trash a couple entries
> in the directory then create valid . and .. files.  I don't know if the
> kernel assumes they're the first two files.  It might be a good idea to
> assume it does.

Would someone please tell me how to use fsdb to do this. I am sitting
with a system with about 10 directories that have been trashed.
The manual has no indication of how to use fsdb. Is there any books that
have more detailed information that I can read? 

Thanks.

Eric Beser
Westinghouse ESG
arpa : ebeser at ajpo.sei.cmu.edu
usenet:eric at sarin.UUCP
(301)-765-1360



More information about the Comp.unix.microport mailing list