Disk
                  File Optimizer frequently asked questions | 
               
             
             Before contacting Hewlett-Packard Company for a problem you are
              having with Disk File Optimizer software, please read through the
              Frequently Asked Questions (FAQs) as well as the product release
              notes. In many cases, you will find the information you need.
              Answers to Frequently Asked Questions
              If the system crashes, will DFO corrupt files?  
               
              No. DFO uses only published OpenVMS interfaces to the file system
              to accomplish its tasks. Data integrity is guaranteed by these interfaces.
              Further, the file optimizer never modifies nor accesses user data.
              If an ongoing defragmentation operation is interrupted by a system
              crash, the volume may be left in a non-optimal state; that is, files
              may not be as defragmented as they could be and they may not be
              placed where they would have been placed had the file optimizer
              completed. However, when the system is brought back up, files on
              that volume are accessible and will not have suffered any integrity
              problems. 
              Does DFO change the modification date on the file?  
               
              No. No dates in the file header are changed. 
              Will DFO run on my database disk (Rdb for example)?  
               
              Yes. If the database is open, DFO does not defragment the database
              file. If the database is closed, DFO treats the database file as
              any other file. 
              How do I check if the scheduler is running?  
               
              Use the following command to check if the scheduler is running on
              a node: 
             $ SHOW SYSTEM/OUTPUT=X.X $ SEARCH X.X DFG 
             If the scheduler is running, you will see DFG$node_name as shown
              below: 
             218002A8 DFG$LUNTIC LEF 6 4091 0 00:01:03.21 1263 1781 
              How do I monitor the progress of a defragmentation process?  
               
              You can monitor the progress of a defragmentation job using the
              following command: 
             $ DEFRAGMENT MONITOR script_name or volume_name: 
              Can I defragment a RAID array?    
              Yes. A controller-created RAID array appears as a normal, but possible
              very large, logical volume and is defragmented as such. The logical
              volumes presented by the StorageWorks RAID Software for OpenVMS
              or POLYCENTER Disk Striping from RAID 5 arrays or stripesets may
              also be defragmented. When you create a script, specify the RAID
              array. Do not defragment individual members of the RAID array as
              they consist of one large file that is not fragmented and should
              not be moved. 
              Can I defragment a shadow set?    
              Yes. When you create a script, specify the DSAx: volume. Do not
              defragment individual members of the shadow set. 
              How do I know when it is time to defragment INDEXF.SYS.  
               
              There are two indications that INDEXF.SYS needs defragmenting: 
             
              -  If you receive a "file header full" error, you should defragment
                INDEXF.SYS. This error means that INDEXF.SYS cannot extend any
                more without the file header extending into another header. The
                file header cannot have more than one header.  
 
              -  If INDEXF.SYS has more than five extents, it is a candidate
                for defragmentation. By OpenVMS file system design, INDEXF.SYS
                has a minimum of four extents. By DFO product design, DFO will
                only defragment INDEXF.SYS if it has more than five extents. 
              
 
             
            Use the following command to check the number of extents in INDEXF.SYS: 
             $ DEFRAGMENT SHOW volume_name:/VOLUME/LOCATION=[000000]INDEXF.SYS 
              Can you defragment INDEXF.SYS on a system boot volume?  
               
              No. You have to dismount the disk before you can defragment INDEXF.SYS. 
              How long should I expect a defragmentation operation to take?  
               
              The time for a defragmentation operation to run depends on the following
              factors: 
             
              - How fragmented the volume is  
 
              -  How much free space is available  
 
              - How busy the volume is with file creation and deletion 
              
 
             
            Many volumes can be defragmented in less than an hour. Some volumes
              can take more than 24 hours. The more the volume is fragmented,
              the more work DFO has to do. If this the first time the volume is
              defragmented, the defragmentation operation could take a long time.
              If the defragment processes are scheduled regularly, they will take
              less time. 
              How can I truncate directories?    
              Some users have directories with large allocated unused blocks after
              files have been deleted. This results in the need to truncate these
              directories. DFO does not provide this feature because it is already
              available using the OpenVMS DCL command SET FILE/TRUNCATE file_name.DIR. 
             The following shows an example using the SET FILE/TRUNCATE command
              to truncate a directory by deleting the unused blocks: 
             $ DIRECTORY/SIZE=ALL DIR/SIZE=ALL [SLICK]MAIL.DIR 
             Directory DSA422:[SLICK] 
             MAIL.DIR;1 115/140 
             $ SET FILE/TRUNCATE [SLICK]MAIL.DIR $ DIR/SIZE=ALL [SLICK]MAIL.DIR 
             Directory DSA422:[SLICK] 
             MAIL.DIR;1 115/116 
             Total of 1 file, 115/116 blocks. 
             Notice in the above example that the directory is 140 blocks in
              size but the end of file is at block 115. The directory is then
              truncated to 116 blocks with the SET FILE /TRUNCATE command. 
              
   
		 |