It sounds like the inbound ftp process is looking for a file it's not
finding, and is getting stuck trying to open a file that doesn't actually
exist. Possibly that file was just renamed, and the directory entry hasn't
actually been updated by the file manager when the next step in the process
attempts to get the newly renamed file, so it locks an internal resource
waiting for the file to open that's supposed to have been renamed by then.
You've most likely got a timing issue.

The reason any subsequent process that tries to do anything with the file
gets locked too is it just joins the list of processes waiting on that
resource, at the end of the line.

Change the ftp process to rename the file, pause for a period of time, then
'get' the newly renamed file. The file manager has to be given time to do
housekeeping, and the ftp process is jumping in too quickly to grab the
newly renamed file.

When you do steps like this manually, the pause in time to type commands is
more than enough for the file manager to complete the directory updates.
When it's done in a job though, the time between 'rename' and 'get' can be a
few clock cycles; enough for the ftp process to issue the rename command,
the file manager to start down it's process tree to complete the rename, but
it doesn't complete in time for the ftp process issuing the next request to
'get' the newly renamed file and it attempts to grab the file before the
file manager has finished.

Stick in a pause for 2-3 seconds between 'rename' and 'get'. A 2-3 second
pause is a very small price to pay in time to not have to reboot your
machine or cause problems with your backups.

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *