[comp.os.vms] Backing up the system disk from DCL

abbonizi@NADC.ARPA (M. Abbonizio) (08/03/88)

Hello VMS managers,
    
   Does anyone backup their system disk from the DCL command level?  What are
   the drawbacks?  Do you use /IGNORE = (INTERLOCK or NOBACKUP)?  What effects
   does this have on PAGEFILE.SYS and SWAPFILE.SYS which are marked "No Backup"
   when restored?  Any info or hints appreciated.

   The system is a cluster running VAX/VMS 4.5.

      Mike Abbonizio

      ARPAnet - abbonizi@NADC.arpa

helen@uhccux.uhcc.hawaii.edu (Helen Rapozo) (08/07/88)

In article <8808031554.AA11042@NADC.ARPA> abbonizi@NADC.ARPA (M. Abbonizio) writes:
>Hello VMS managers,
>    
>   Does anyone backup their system disk from the DCL command level?  What are
>   the drawbacks?  Do you use /IGNORE = (INTERLOCK or NOBACKUP)?  What effects
>   does this have on PAGEFILE.SYS and SWAPFILE.SYS which are marked "No Backup"
>   when restored?  Any info or hints appreciated.
>

We use /IGNORE=INTERLOCK when we backup the system disk, but never with
NOBACKUP.  When any file is encountered with NOBACKUP it will only
backup the file header information like how big the file is and other
attribubes.  That way when you restore the file it will take up
that much space on the disk.  The contents of the file will be
meaningless but these files only contain temporary information
anyway.

Just to let you know that while we back up our system disk fully
every week, we still do a standalone backup of the system drive
once a month (in place of a weekly).

We run a VAX 8200 and two MicroVAX II.

-- 
BITNET address: helen@uhccux
UH Admin Network: HCCADA::CS_RAPOZO
US Mail Address: 874 Dillingham Blvd., Honolulu, HI 96817, Ph# (808) 845-9202
PLATO address: helen rapozo/honcc/hawaii

GG.SPY@ISUMVS.BITNET ("John Hascall") (08/08/88)

> Date:         Wed, 3 Aug 88 11:54:54 EDT
> From:         "M. Abbonizio" <abbonizi@NADC.ARPA>
> Subject:      Backing up the system disk from DCL
> To:           John Hascall <GA.JPH@ISUMVS>
>
> Hello VMS managers,
>
>    Does anyone backup their system disk from the DCL command level?  What are
>    the drawbacks?  Do you use /IGNORE = (INTERLOCK or NOBACKUP)?  What effects
>    does this have on PAGEFILE.SYS and SWAPFILE.SYS which are marked "No Backup
"
>    when restored?  Any info or hints appreciated.
>
>    The system is a cluster running VAX/VMS 4.5.
>
>       Mike Abbonizio
>

   I use:

      $ backup /buffers=5 /ignore=interlock /image /record /rewind -
         'disk'  'save_set'/LABEL='backup$volser'

   (Note that this is in a command procedure, hence the symbols...)
    It just gives a message that the files are marked NOBACKUP and
   continues on it's merry way.

     While on the subject of backup  (* FLAME ON *)  One thing which
  annoys me to no end is that when you specify /record (to record the
  time of the backup in the file header) and a user deletes a file after
  the time it is backed up and the time the file header is to be updated
  backup gets all bent out of shape...It would certainly be nice if this
  was handled as nicely as /nobackup files are..

John Hascall
ISU Comp Center

BRUCE@LOCKE.HS.WASHINGTON.EDU (BRUCE EDWARDS) (08/08/88)

Mike,

     We tend to back up our system disk at DCL level.  We use /IGNORE=INTER
and the PAGEFILE and SWAPFILES get created on a restore, but without any
data.  Our system disk is an RA81 and we generate an RP06 offline pack with
the system root every so often.  After we restore the disk from the Backup
tapes we install the additional pointers for the [SYSEXE] and [SYSMAINT]
directories write the Boot block and use an different boot file to bring
up the system.  We also backup (to another disk) every hour or so some
key datafiles.  Things may not be perfect when you restore, but you tend
to lose less.  We also keep an eye on which files are interlocked to spot
any new problems.  It is a lot easier to set up this procedure ahead of
time then to wait until problems occur.  This way we can run in a degraded
mode until the Disk is repaired and also allows us to regenerate the system
disk on a new RA81.  We have run this procedure with both HSC and UDA based
RA-81's and are currently running VMS 4.7.  The major drawbacks relate to
what files are open.  Some files need to be individually repaired,
some we just grab a different copy and some we ignore.  We keep 3 backups
of our hourly important data file backup. Hope this helps

Bruce Edwards
John Locke Computer Center
University Of Washington SB-50
Seattle, WA 98195

Bitnet  -- BRUCE@UWALOCKE
Internet--bruce@locke.hs.washington.edu (128.208.10.1)