jw at pegasys.ws
Fri May 31 17:36:02 EST 2002
On Fri, May 31, 2002 at 06:27:27PM +0200, Dick Streefland wrote:
> On Thursday 2002-05-30 12:43, Michael Montero wrote:
> | That's great news. I believe this applies to me just fine and I
> | can turn off the checksum. Quick question....can anyone explain to me
> | when the data in a file might change without changing the mtime, ctime or
> | size? I'm not sure I've ever come across that before. An example might
> | help me determine if I can safely remove -c.
> I ran into a situation where I had to use the -c option. I use two
> rsync commands to install Linux onto new PCs, the first one to "clone"
> a system, and a second one to update a few host specific files such as
> /etc/hostname, /etc/network/interfaces etc. The host specific files
> are automatically generated from template files, using a compact
> description of each machine. This script runs so fast that multiple
> files will get the same mtime. So, it is possible that some host
> specific files are not updated by the second rsync command, because
> the mtime is the same as the mtime on the master.
> On one occasion, a machine came up with the same hostname as the
> master machine after the installation. Furtunately, the file with the
> IP number was correct! I now use the -c option on the second rsync,
> which is not a problem, because the number of files is very limited.
> BTW: I previously used NFS with find/cpio to clone Linux installations,
> but rsync is *way* faster, 2-3 times faster on a 10 Mb network.
Interesting story and a good example.
PS. I would have inserted a delay between generating the
master set and the machine specific ones or recursively touched
the sets with different times (find blah -type f -print
|xargs touch). Much faster than -c.
J.W. Schultz Pegasystems Technologies
email address: jw at pegasys.ws
Remember Cernan and Schmitt
More information about the rsync