Steve Bertrand wrote: > xufengnju wrote: >> Hi all, >> I have a storage server that holds one million of images with well >> structured directory structure such as >> >> /data/user1/2008/09/12/image1.jpg >> /data/user1/2008/09/12/image2.jpg >> ... >> /data/user2/2009/01/01/image1.jpg >> ... >> >> I want to copy them to the /data2 directory in the same server on another >> disk partion. >> I want to keep the directory structure and `chown` && `chmod` the >> directories and files,much like doing a `cp -rf /data /data2 && chown -R >> sysuser:sysuser /data2 && chmod -R 755 /data2`. >> >> File::Find maybe an option. >> Is there somebody who have some suggestions? >> >> If I do a `cp -rf /data /data2 && chown -R sysuser:sysuser /data2 && chmod >> -R 755 /data2`,how much time maybe taken to finish the job?(The images are >> about one million in count and 250GB in size totally). > > Perhaps I am missing something completely obvious. If not: > > Why-oh-why do you want to use Perl do perform such a task? > > Use dump/restore (which I can't recall a cli sequence for off the top of > my head),
...and the dump commands I had to look up. Use it at your own risk. Essentially, it creates a full (0) dump of /data, and because of '-f', writes it to '-' (or STDOUT). It is subsequently piped to the process on the right, which changes directory to your backup disk, /mnt/data, and then pristinely restores the incoming dump into '-' ( the current file system, /mnt/data ). # mount your backup disk to /mnt/data % dump -0f - /data | ( cd /mnt/data; restore -rf - ) Perhaps there is a Perl way to do it, but otherwise, for 250GB of data, research dump/restore, and test it out (after making a backup). imho, you shouldn't use another layer of abstraction for managing such a large volume of data, unless you are attempting to create some sort of index for it. Steve
smime.p7s
Description: S/MIME Cryptographic Signature
