Disk Cluster Size too Large

[Andy Bustamante]

You can use the BACKUP /TRUNCATE switch in to truncate files at [EOF] when restoring to your newly initialized disk. The default behavior of BACKUP will be to restore the file with the previous allocation.

Andy

[Karen Lee_3]

the backup/truncate switch doesn't seem to have any affect - the files are still maintaining the 1372 allocation.

[Uwe Zessin]

Did you use BACKUP/NOINITIALIZE/TRUNCATE to restore to a disk that was INITIALIZEd with a smaller cluster factor?

[Jan van den Ende]

Karen,

please check your "new" clustersize.
If it is back to 1372 (which I suspect) then you probably did NOT add /NOINIT to your BACKUP, and that undid YOUR init by re-initialysing the disk with the setting out of the saveset..

Also, please re-read Hein's last question.
Do you have a COMPELLING reason for so large a RAIDset? And by compelling I mean: a single file too big (or expected to grow as much) to fit onto a single drive? In which case, a big clustersize actually would be an advantage.
But for your many small files, there is NO reason to put them on a big device (or RAIDset), but many reasons to use smaller devices. (not the least of those: manageable BACKUP savesets and rstore procedures)
Should you have an applic that mixes both, then it is REALLY worth the effort of implementing clever choices of Logical Names, maybe by using search lists.
Should you need help on that, then please supply some details.

hth

Proost.

Have one on me (maybe in May in Nashua?)

jpe

[Karen Lee_3]

i redid the init and then backkup/truncate command on a smaller raid array and it worked fine - must have been late....

The only reason we have for the very large raid sets is the footprint we have for the cab. with the need for a LOT of small files.

[Jan van den Ende]

Karen


The only reason we have for the very large raid sets is the footprint we have for the cab. with the need for a LOT of small files.


I do not quite get that argument.
On exactly the same footprint, using the same physical drives, it is also possible to present "any" reasonable number of smaller units, with the same total storage.

fwiw

Proost.

Have one on me (maybe in May in Nashua?)

jpe