it's not completely clear to me how this filesystem is used and how you are 
planning to manage your alternate service, so the following may or may not be 
useful...

-- unison is similar to rsync but is bidirectional and xfers can beinitiated 
from either side.  so if you will need to repeatedly update the offsite fs from 
the onsite fs and then reverse the process, this could simplify your scripting 
a bit. 

-- openafs also has agresive persistent client side caching and may provide the 
functionality you were hoping to get from cachefs.  i'm not sure about this, 
though and at the moment their web site's not responding, at least from here.  
[ always a bit disturbing when a network project''s network doesn't work :-) ]  

-- the special magic of rsync and unison is that when used to keep changing 
filesystems in sync they detect the changes and xfer only whats needed to keep 
things the same.  if your moving everything from one place to another a blunter 
instrument might do the job ok.   rdist or tar for example (see the tar man 
page for an example of replicating directory contents over a network.) 

-- note that all of the above can be tunneled over ssh if need be.

-- though shooting in the semi-darkness it sounds like you might arrive at a  
convenient  arrangement by using a combination of unison for managing the data 
and nfs failover for managing the mounts.   however, you need to be careful to 
keep the unauthoritative copy of the data offline so clients will fail over to 
the authoritative one.  

-- an alternative to the above is to use rsync for everything.  you can use it 
for shuttling the data back and forth,  that's obvious.  assuming you have root 
all around, you can also use it push updated copies of /etc/fstab to the 
clients.   that way you can move the data, verify that it's reachable, and 
point everybody to it.  (if you're using a directory service the same concept 
applies, but you'd do it differently.)
 
 
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to