[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::digital_unix

Title:DIGITAL UNIX(FORMERLY KNOWN AS DEC OSF/1)
Notice:Welcome to the Digital UNIX Conference
Moderator:SMURF::DENHAM
Created:Thu Mar 16 1995
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:10068
Total number of notes:35879

8721.0. "copy 60GB of raw data??" by DYOSW5::WILDER (Does virtual reality get swapped?) Wed Feb 05 1997 08:09

    What is the best way to make a mirror copy of a raw, 60GB database? dd
    would work, but there is not much error checking with it, and we need
    to make sure that the data on the receiving disks IS GOOD, as they will
    use this same method to "rebuild" there database for testing purposes.
    
    Any suggestions?
    
    Thanks,
    
    /jim
    
T.RTitleUserPersonal
Name
DateLines
8721.1XIRTLU::schottEric R. Schott USG Product ManagementWed Feb 05 1997 12:349
Use the database backup program?

Use LSM to make a mirror, and split the mirror?

I thought dd does do error checks, but not comparison...

Write your own program?


8721.2IOSG::MARSHALLWed Feb 05 1997 13:127
Call me hopelessly na�ve, but:

    cp /dev/disk1 /dev/disk2
    cmp /dev/disk1 /dev/disk2
?

Scott
8721.3dd then cmp?QUARRY::reevesJon Reeves, UNIX compiler groupWed Feb 05 1997 13:145
Assuming the database is quiet at the time of the copy, you can use cmp after
the dd to confirm the copy is identical.

If the database isn't quiet, then you'll definitely need a vendor-specific
copy program.
8721.4NABETH::alanDr. File System's Home for Wayward Inodes.Wed Feb 05 1997 17:275
	dd(1) will probably be a better choice than cp(1) for making
	the copy, because dd(1) will use the transfer size of your
	choice.  Since raw disks will treat a long read (beyond the
	end of a device) as an error, getting the right transfer
	size could be important.
8721.5dd the LSM raw volume??DYOSW5::WILDERDoes virtual reality get swapped?Sun Feb 09 1997 17:009
    Okay, the general consensus seems to be dd. Now, the 60GD database is
    really 18 4GB spindles with 46 LSM stripped volumes. Now, if I use dd,
    do I pass dd the LSM raw volume name (/dev/rvol/....)? Also, since
    these are raw volumes, will the default dd blocking factor be okay?
    
    Thanks,
    
    /jim