[comp.os.vms] Disk defragmenters.

MACALLSTR@vax1.physics.oxford.ac.UK (06/29/88)

While disk defragmenters may only produce dramatic performance improvements
 on very badly fragmented disks ( usually disks with little free space ) for
 individual applications, you will a notice drop in BACKUP time after the
 disk has been defragmented. 
I have tried DISKEEPER and RABBIT-7 and am using RABBIT-7 mainly because at
 the time the decision was made ( about a year ago ) RABBIT-7 gave more
 flexibility. Anyone using disk defragmenting packages ought to be aware of
 what the package is doing to the arrangement of files on the disk ; free
 space may be concentrated in one chunk at the beginning/end/somewhere in 
 between of the disk or it may be divided into several separate chunks
 distributed over the disk. You must discover which formula suits your system :
 do you have many small/medium/large files or does the disk contain several
 large database files for example? 
John

brent@uwovax.uwo.ca (06/30/88)

In article <8806282211.AA01437@ucbvax.Berkeley.EDU>, MACALLSTR@vax1.physics.oxford.ac.UK writes:
> While disk defragmenters may only produce dramatic performance improvements
>  on very badly fragmented disks ( usually disks with little free space ) for
>  individual applications, you will a notice drop in BACKUP time after the
>  disk has been defragmented. 
> I have tried DISKEEPER and RABBIT-7 and am using RABBIT-7 mainly because at
>  the time the decision was made ( about a year ago ) RABBIT-7 gave more
>  flexibility.    <stuff omitted here to reduce posting size>
> 
   Funny.  We looked at SQUEEZPAK (DEMAC) and RABBIT-7.  (DISKEEPER wanted up
front money to evaluate, so it wasn't. :-)  Both did about the same job, but
we had better results from RABBIT-7 on a severely fragamented disk (which was
also *very* full.  Despite the fact that this is not the normal way we'll use
the s/w (in ongoing runs, they came out about the same), there was a slightly
cosier feeling about RABBIT.  And when pressed, both companies' pricing became
very similar.  BTW the file *placement* facility in v3.0 looks really slick.
Once we get it up and running, the net may get a review.  Cheers.  Brent.