Help understanding rsync and cwrsync
dlochart at gmail.com
Thu Mar 2 03:51:16 GMT 2006
On 3/1/06, Peter <petermatulis at yahoo.ca> wrote:
> --- Doug Lochart <dlochart at gmail.com> wrote:
> > Windows (cw)Rsync Client --> Linux Rsync server
> > 1) Should we abandon cwrsync for cygwin + rsync?
> > Have the
> > disparities between the two versions been resolved?
> I am using cwrsync and it runs fine connecting to
> Slackware server.
I asked because in the mailing list archives I found this
"I want to make sure that cwRsync is NOT a separate product/solution or
whatever. It is simply a packaging of rsync on cygwin to make
deployment/installation easier. cwRsync is a minimalist environment and does
not support keeping ownership information between transfers."
It was posted by a cwrsync developer. We were having issues with file
permissions not being transferred so this post caught my attention.
It was dated in 2004 so I was wondering if this limitation still
existes. I am not saying we did nothing wrong but it looks right and
then I saw this post. So cwrsync works fine for you and you get all
your permissions etc?
> > 2) What are the risks with using cwrsync (we want a
> > limited footprint
> > on the client)?
> What risks? The "footprint" is small.
I meant risks of cwrsync vs rsync over a full blown cygwin
installation. Again I am asking this based on the post I found. I
know the cwrsync footprint is small but a cygwn installation with
rsync would be bigger.
> > 3) Assuming we use straight cygwin + rsync how does
> > the linux rsync
> > server handle the permissions and ACL's of an NTFS
> > files system? Are
> > they restorable back to a windows machine? The
> > linux server won't
> > have any of the users plus the ACLS stuff is totally
> > different that
> > unix.
> Can't say. I don't use cygwin.
Ok, how does cwrsync handle ACLs, users and permissions as on the
windows box? How/where is that info kept in the filesystem on the
linux box? Is it in meta-data? I figure basic file permissions are
stored with the file but how is the user stored? I can't imagine we
would have to create a user account for that end user ... or would we?
> > We will be using rsync over an SSH tunnel to start
> > but will eventually
> > switch to rsync over SSL.
> Why? What is wrong with SSH? I do not believe rsync
> supports SSL natively yet.
This will be deployed hopefully to many people. We feel that an SSL
certificate would be easier than setting up seperate keys for each end
user. The rsyncs would happen in nightly jobs without user
> > We want to be able to add module definitions to the
> > rsyncd.conf file
> > as we go along without having to restart the rsync
> > daemon or inetd
> > service. We want this so that we can add new client
> > machines without
> > interfering with other clients use of the server.
> > Is what we are
> > trying to do workable? If so how should we run our
> > rsync server.
> The config file containing the module settings is read
> by rsync at every connection so you do not need to
> restart the daemon. Alternatively, you can run rsyncd
> on a single-transfer basis. The remote user invokes
> the daemon.
Could you explain how the last part works? (the remote user invokes
Most examples I have seen mention running it in daemon mode or via inetd?
Thanks again for your insite
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
What profits a man if he gains the whole world yet loses his soul?
More information about the rsync