Copy 300 gb data

[7280]

Ciao,
I want to move 300gb data from a nas storage to a san.
Server is Redhat EL AS 3 upd 3. Data are domino archive db's.
What is the fastest way to do this?
Any suggestion would be appreciated.
Unfortunately I don't have any backup tool like TSM.

Thanks,
Tarek

 

[7280]

Please give me some ideas.
I thought about two solutions:
1) cp -rp /old_dir /new_dir

2) cd /old_dir
find . -depth | cpio -pdvm /newdir

thanks,
Tarek

 

[7280]

Hi,
last thing.
File systems are different.
/olddir is nfs
/newdir is ext3

Thanks again.

 

[Trevoke]

No very fast way to do this, but I do wonder if reiserfs would not be a better newdir choice. Not sure though, it's your data

I think the two ideas you had are likely to be the best ones out there... Make sure you have a very fast connection between the two.

"That time in Seattle... was a nightmare. I came out of it dead broke, without a house, without anything except a girlfriend and a knowledge of UNIX."
"Well, that's something," Avi says. "Normally those two are mutually exclusive."
-- Neal Stephenson, "Cryptonomicon"

 

[7280]

Hi,
cp seem finished, I can't see process with ps.
But dimensions of two directories are different.
olddir: 306 gb
newdir: 246 gb

But number of files are the same and in the log file there's nothing written.
How is it possible that space used is different?
Is it possible that during copy a sort of defragmentation is done?
Is it possible that same file occupies different space on ext3 fs insted of nfs?

I want to be sure that my copy was fine but I don't know how!!
Please suggest, it's very urgent!

Thanks,
Tarek

 

[7280]

Hi,
du shows almost same usage:
du -sm olddir: 251253
du -sm newdir: 251559

but with df -m I see:
olddir: 312397
newdir: 251591
Thanks again,
Tarek

 

[motoslide]

If the files are currently static, you can run a checksum on the source and destination, then compare the results.

Something like this:
cd /old_dir
find . -type f |sort|xargs sum > /tmp/old.sumlist
cd /new_dir
find . -type f |sort|xargs sum > /tmp/new.sumlist
cd /tmp
diff old.sumlist new.sumlist >diff_file

Note that the sum command will be slow on very large files.





"Proof that there is intelligent life in Oregon. Well, Life anyway.