Also a long time since I used VMS. I do remember that one issue I sometimes faced is that the free block count on a disk device would sometimes not match up with reality (might be the issue you face). To correct this we would use copy/allocate command... something like... $ copy/allocate=nnnnnn junk.dat junk2.dat where junk.dat was a small file and "nnnnnn" was some large number to force the command to fail. This had the side effect of resetting the free block count for the device. Regards, Ken. Business Applications, Solutions and Service Delivery (BASSD) Database Services, Howard Street, New London, CT Office: (860) 732-1422 | Mobile: (860) 389-3534 ________________________________ From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Andrew Kerber Sent: Wednesday, September 01, 2010 1:41 PM To: ian@xxxxxxxxxxxxxxxxx Cc: Oracle-L Freelists Subject: Re: Failure to Release Diskspace Dont you have to kill the process accessing the file before it can be resized? show dev/files and kill the processes that is accessing the file? Its been a few years since I used VMS myself... On Wed, Sep 1, 2010 at 12:31 PM, MacGregor, Ian A. <ian@xxxxxxxxxxxxxxxxx> wrote: I have not touched a VMS system for 20 odd years. One of the things I did was to shrink the size of the undo tablespace from 2770 MB to 1024 MB. When I issue the VMS command show device/full <device_name> the number of free blocks is no different than before. What needs to be done to get the space released. Ian MacGregor SLAC National Accelerator Center -- //www.freelists.org/webpage/oracle-l -- Andrew W. Kerber 'If at first you dont succeed, dont take up skydiving.'