[comp.os.vms] Another Viewpoint On Volume Shadowing And Other Areas Of Interest...

CLAYTON@XRT.UPENN.EDU ("Clayton, Paul D.") (08/20/87)

Information From TSO Financial - The Saga Continues...
Chapter 17 - August 9, 1987

Jim Dibble has requested the following information concerning the use of
Volume Shadowing and its impact to the systems.
	 Specific questions:
		What hardware / software versions
		How many volumes shadowed, types of drives, system, non-system
		Performance changes (+ or -) actually seen
		Volume backup procedures for system and non-system disks
		Problems

Celeste Finison responded with a comprehensive answer but I wish to add some
comments to what she outlined due to our installation of the product.

!	If you are currently running VMS v4.5 you can skip the step in
!	the installation procedure that requires you to create a new
!	VMB.exe.  You already have it.

My understanding is that the new VMB.EXE that is needed on the boot media 
(Floppy/PRO) is in different states depending on the system. If you have a 
8XXX with the PRO console at Version 4.0 or higher then the new VMB.EXE does
NOT have to be copied. In all other cases the new VMB.EXE does need to be 
copied.

The following is the last line and concerns the speed of the disk copy.

!	This will get the laspsed disk and perform a full copy operation 
!	in approx. 15-20 minutes.

I have an HSC70/HSC50 combination that is used to support RA/SI disks. Last
week when I added a RA81 disk to the shadow set, the disk copy operation which
takes place in the background, took 2+ hours to complete. This was done on the
HSC70 and it was heavily loaded at the time. The 15-20 minute deal is with 
little or no activity.

There are two warnings I have. First, dismounting a volume from a shadow set 
does NOT trigger a flush of all process buffers which are used for files that 
are open on the volume at the time of the dismount. The result of this is that
the files themselves, especially if they are ISAM files, can be in a 'corrupt' 
state due to updated buffers containing buckets that have yet to reach the 
disk. The bottom line is that this ability to dismount a shadowed disk anytime 
does NOT really help with backups unless you can get a short period of time 
when the disk has no open files for update.

The second deals with having multi HSC's and the disks split between them. The
maximum number of shadow sets per HSC under V1.0 is 4 and V1.1 allows for 7.
Should the HSC's/disks be configured in such a way that under HSC failover
conditions more then the allowed number of shadow sets are 'forced' onto the
remaining HSC(s). The ones over the maximum go to BIT BUCKET LAND, and you
have NO control over which ones they are that go away, its a case of last
set(s) failed are the losers.

Other then the above, we have the same type of improvment is access times that
Celeste has indicated. We do not shadow the system disks. The bulk of our I/O
is done on data disks and that is what is shadowed.

Paul D. Clayton - Manager Of Systems
TSO Financial - Horsham, Pa. USA
Address - CLAYTON%XRT@CIS.UPENN.EDU