[Info-vax] RAID vs. MOUNT/BIND

[Hein RMS van den Heuvel]

On Nov 17, 8:19 am, koeh... at eisner.nospam.encompasserve.org (Bob
Koehler) wrote:
> In article <ic0cpb$6h... at online.de>, hel... at astro.multiCLOTHESvax.de (Phillip Helbig---undress to reply) writes:
> >  If the only point is to create a larger logical disk, which is better?
"
>    Volume sets can spread files over two physical disks and make them
>    look like one, but cannot spread a single file over two disks.  

Yes it can. It will just create an extension header on an other
member.
This can be need (no more blocks, as per example below ) or speed
(design) based.

You can design the split using VOLUME PLACEMENT in the FDL when
creating an indexed file and you can control this during SYS$EXTEND or
QIO EXTEND calls.

STRIPING (RAID-5) of course has the potential advantage of equally
spreading IO, but that was stated as a non-goal.

Hein


$ copy /all=4500 nl: disk$lda_10_11_12:[test]split.dat
$ show dev disk$lda_10_11_12
EISNER$LDA10:           Mounted alloc        0  LDA10
805     1   1
EISNER$LDA11:           Mounted alloc        0  LDA11
0     1   1
EISNER$LDA12:           Mounted alloc        0  LDA12
0     1   1
$ dir/file/size=all disk$lda_10_11_12:[test]split.dat
SPLIT.DAT;1          (12,1,3)                      0/4500

$ pipe dump/head/block=count=0 disk$lda_10_11_12:[test]split.dat |
search sys$pipe " id ",lbn
File ID (12,1,3)   End of file block 0 / Allocated 4500
        Count:       1869        LBN:        131
File ID (12,1,3)   End of file block 0 / Allocated 4500
        Count:       1768        LBN:        232
File ID (12,1,3)   End of file block 0 / Allocated 4500
        Count:        863        LBN:        332