ctdb-scripts: Use more unique temporary file names

ctdb-scripts: Use more unique temporary file names

Consider this sequence of events:

1. Instance of script running update_tickles() hangs
2. Script debugging is launched asynchronously
3. New instance of script is launched, creates temporary file(s)
4. Original hung script makes progress before asynchronous script
   debugging kills it, so it removes temporary file(s)
5. New instance of script produces error due to missing files(s)

This is obviously rare.

Use more unique filenames to avoid step (4) removing the file(s)
belonging to other instances of the script.

This requires some extra cleanup to avoid too many temporary files
(which is why unique filenames were not originally usd).  It is
sufficient to remove files modified at least 10 minutes ago.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Michael Adam <obnox@samba.org>