Operating System - Tru64 Unix
1748104 Members
4458 Online
108758 Solutions
New Discussion юеВ

Re: copying 3.5tb of data

 
anthony m. gonzales
Frequent Advisor

copying 3.5tb of data

Hello,

I am getting ready to embark on copying 3.5tb of data and was wondering what would be the recommended way to do this. I do not have all the details, but I believe I will have disks that are identical so dd might be an option. I do get confused as if needed to restore the data from a dd copy. I know that reversing the command used to dd the disk, but, is the disk now ready to mount and see all the files?

Thanks,
Tony
9 REPLIES 9
Rob Leadbeater
Honored Contributor

Re: copying 3.5tb of data

Hi Tony,

Can you provide us more information please ?

What are the source and target storage systems ?

What OS ?

What servers ?

Cheers,

Rob
anthony m. gonzales
Frequent Advisor

Re: copying 3.5tb of data

They are HSG80 DISKS, on TRU64 UNIX 5.1B, ES45...That is all I have to work with for now..

Tony
Steven Schweda
Honored Contributor

Re: copying 3.5tb of data

> [...] dd might be an option. [...]

Many people seem to like "dd" for a job like
this, but I normally prefer something like
a vdump|vrestore pipeline.

> [...] is the disk now ready to mount and
> see all the files?

"dd" may do everything right, and I may worry
too much, but I always worry about what might
go wrong if the (logical) disk size is
different, what happens when there are two
identical disks on the system, and so on.
Besides, I'd expect vdump|vrestore to
defragment everything, and that seems like a
good thing.

Of course, I've never needed to wait while
3.5TB moved through the pipe, and "dd" may be
faster.
Rob Leadbeater
Honored Contributor

Re: copying 3.5tb of data

Hi,

Another option, if you have the AdvFS Utilities license, is to use addvol and rmvol.

Add a new volume from the target storage system to your AdvFS domain with addvol, then remove the source system's volume with rmvol.

This will copy all the data, whilst still allowing access to the file system. I wouldn't like to say how long 3.5TB would take though...

Cheers,

Rob
anthony m. gonzales
Frequent Advisor

Re: copying 3.5tb of data

Rob,

Thanks for your reply. I had not considered that. Food for thought. If I do use dd I know that I would reverse the original dd command, but, will all the "new" disks then just mount as though they were filesystems or can I actully just do a mount /dev and view the data prior to releasing the system.

Thanks,
Anthony
Rob Leadbeater
Honored Contributor

Re: copying 3.5tb of data

Pass... I've never used dd to copy file systems.

As Steven suggested, a pipelined vdump|vrestore would be my preferred method, if you can't go down the addvol/rmvol route...

# vdump -0f - /source | vrestore -xf - -D /destination

Cheers,

Rob
anthony m. gonzales
Frequent Advisor

Re: copying 3.5tb of data

Rob,

Thanks for the answer...I am just looking for the most efficient way to backup/restore the data. I have no idea the time it would take to backup/restore 3.5tb...I am thinking 24 hrs...who knows...

From what I read dd seems to work, but, I have not seen or heard of anyone bringing a system up and see the dd data as true files...

Thanks,
Anthony
DCBrown(2)
New Member

Re: copying 3.5tb of data

dd will create an exact copy of the data... but depending on the file system it might not mount too well/easily. If memory serves, I think it was ADVfs that included the device wwid in the metadata so that it could determine if it was looking at the original filesystem or a cloned copy.

Better off using the backup/restore method recommended for the filesystem type you are using.



anthony m. gonzales
Frequent Advisor

Re: copying 3.5tb of data

DC,

Thanks for your input on this. As I have mentioned it is 3.5tb of data and I was looking for the fastest most efficient way to do this. I know that fast does not translate to efficient sometimes. The file systems are all ADVFS...

Thanks,
Anthony