Yesterday I wrote regarding an issue we had where removing a VOB
failed because of .nfsXXXX files created in the VOB storage directory
while it was in the process of being removed.
In this case, we have Sun servers interfacing with a Network Appliance
filer.
The problem is that while I was umounting the VOB before issuing the
rmvob command, I was not searching for the vob_server process and
killing it (e.g., kill -15) before issuing the rmvob command.
By killing it before removing the VOB, the creation of .nfsXXXX files,
which occurs when an open file is removed by an NFS client, is
avoided and the rmvob completes normally.
I guess I'm surprised that rmview works OK -- perhaps it kills the
view server process, while rmvob does not.
Recognizing that in the past VOBs have been expected to have their
databases on local (not NFS-mounted) storage, while views have not,
would help explain why views don't have this problem, while VOBs
do.
Any reason that a "cleartool umount /vob/x" should not contact
the vob server process for /vob/x and tell it to exit?
John Sambrook
__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
This archive was generated by hypermail 2b29 : Tue Jul 31 2001 - 22:04:01 EDT