It's pretty clear that it is a bug in the kernel code and to be more
precise the cache entry in the dcache doesn't get invalidated apparently. 

Did you mount volumes or anything like it?  This could well be at the root
of the confusion.

Peter


On Wed, 29 Apr 1998, the prophet wrote:

> The problem gets weirder.  I got rid of the rm problem by renaming another
> file to the same name as the first and then removing it.  But now:
> 
> 
> [triffid:1:25pm:0]/coda > ls
> users/
> [triffid:1:25pm:0]/coda > cat > testfile
> testfile: File exists.
> [triffid:1:25pm:1]/coda > test -f testfile || echo not there
> not there
> 
> I wrote little programs to open the file O_RDONLY, O_WRONLY, and O_RDWR.
> In each case, open(2) returned -1 with errno set to ENOENT.  So how cat
> thinks the file is there is beyond me.
> 
> 
>       0
>         0
> \  /\ 0
>  \/ o\ 0 
>  /\  /
> /  \/
> 
> 
>