Hi Johan,

> -----Original Message-----
> Hello,
>
> For synchronization purposes we use a locking strategy, which uses the
Posix
> fcntl call to lock regions of a bytestream file in the HFS environment.
> Under some, yet undetermined, circumstances the fcntl call fails (errno =
50
> [ system call error ]), looking at the file with AIFFILELGET reveals that
> the last error was 7114 (Internal HPFLOCK error, FLOD marked invalid).
Does
> anybody know under which circumstances this error occurs ?
>
> Any help would be greatly appreciated
>
> Holland House B.V.
>
> Johan Harmsen
>

This error condition is seen when the file system encounters an unexpected
trap during the manipulation of the File Lock Object Descriptor for a file.
Because we trapped in the middle of manipulating the data structure, its
integrity is now in an unknown state and the data structure is marked
invalid. This prevents any further, lock, unlock, or lock information
requests from being processed (the data structure is made inaccessible).

To work around this problem, you could copy the original file to a new file,
purge the original and rename the new file to the original file's name. This
trick is used because the file system tries to cache recently opened file
information and a simple close and reopen of the original file will likely
use the cached version of the file data structures which will still be
marked invalid. Going through the copy-purge-rename sequence will fool the
file system into allocating new data structures for the file.

In terms of trying to address the underlying cause(s) of the problem, it
would be helpful if you could take a dump of the system when this problem
occurs. By reporting this problem to the HP Response Center and providing a
dump, we can then analyze the state of the FLOD for your file and try to
deduce where the original problem (the cause of the unexpected trap) may
lie.

I hope this helps!

Take Care,
Craig Fairchild
MPE/iX File System Architect