> We use ufsdump to backup our file systems to 8 mm,
> 4mm or DAT 72 tape drives. The default blocking
> factor ufsdump uses is 64 (without b or c option) and
> the problem with this is the amount of data that can
> fit on a tape because of the IRG.

Does the IRG really cause a substantial difference in the capacity?

> My questions are:
> 
> - how can I find out the optimum blocking factor (or
> that is supported) for each tape device
> (8mm/4mm/DAT72)?

I would suppose you'd just go as high as you can if you're trying to optimize 
for space.

> - what are the issues of changing the blocking factor
> to 512 (ufsdump code seems to cap the blocking factor
> to 512)?

Older operating systems might not be able to deal with very large factors.  A 
factor of 126 was commonly used because some old OS's broke at 128.  I don't 
know if anything has a problem there now.

For smaller tapes or smaller files on tape, you're wasting space in the final 
block because it is likely that more of it will be unused with a large block 
size.

Also 'ufsdump' and the tape have their own blocks.  I imagine it's setting them 
to be the same, but I haven't tried to confirm that.  You could reblock with 
'dd' or something even if the ufsdump code were limiting.

ufsdump 0f - /source | dd of=/path/to/tape obs=<blocksize>

You'd probably need to reverse that on restoration, because ufsrestore isn't 
going to know to use a larger blocksize for reading (you'll get an "out of 
space" error).

-- 
Darren
 
 
This message posted from opensolaris.org
_______________________________________________
ufs-discuss mailing list
[email protected]

Reply via email to