What are you listing there? Did you mount that disk?
And did you unmount it before the copy? If not, there is a chance that the original directory content, before the copy, is still in the file system cache.

You appear to be copying slice2 of the first disk to slice2 of the second disk, is it actually there(?).

Where are you looking for your files(?).

Also, check under /dev on your first disk.

You could use "status=progress", to see what's being written. Also you didn't say what the final result of "records in/records out" info.

outch !
More I read yours remarks more I'm not sure of what I'm doing ...
Goal : duplicate source disk (c0t0d0) to destination disk (c0t9d0); these are same type of disk.

So I summarize here information and what i did.
I follow Solaris admin guide.
I restarted from the beginning to be sure.

1°) Destination disk is formatted and inserted into the SUN's slot. Partitions are identical on both disks (I used format).
2°) I constructed a new ufs fs on c0t9d0s2. (I used s2 because I understood, it represents the whole disk; is that true ?)
3°) I labelled destination disk.
4°) Vfstab still only mention source disk (see 1.png)
5°) I created u09 as mounting point. (see 2.png)
6°) I modified Vfstab but I'm wondering about mount point for new slices(see 3.png)
could c0t0d0s3 has same mounting point than c0t9d0s3 or should i write /usr for source disk and /u09/usr for destination disk ?
The goal is still to remove source disk form slot0 and insert destination disk to its place.

Let's close all the ? before launching dd command.
Thank you.
EB

Attached Thumbnails

     

you must not use the same mount point twice (or more times), that cannot work. use /var2 /usr2 or something similar.
you can use dd on slices (like /dev/rdsk/c0t9d0s2), but in that case that slice must not be mounted.
you can use cp or rsync (if you have) on directories - on mounted filesystem.
But you must not umount /var and /usr on a running OS, that will most probably cause a dead system.

You are trying to copy one disk to the other, therefore, use the disk designation, not slices.

https://docs.oracle.com/cd/E19253-01...904/index.html

https://docs.oracle.com/cd/E36784_01...871/dd-1m.html

http://ibgwww.colorado.edu/~lessem/p...aris/dd.1.html

That's an older UNIX device naming convention, especially on Sun/Solaris/SVR4.2 systems: controller, target (as in SCSI target), disk (?), and slice (i.e., partition). If memory serves slice 2 means "entire disk" (as did the "c" partition on DEC UNIX/Tru64). I can't recall ever seeing anything other than "d0" in a Solaris disk device name.

I'd always hoped Linux would adopt it and leave disk letters behind but we got UUIDs instead. Oh, well...