[Csync2] [gaio at sv.lnf.it: First use and first question...]

[Marco Gaiarin]

Mandi! Giampaolo Tomassoni
  In chel di` si favelave...

[...spero sopravviverai al mio pessimo inglese... ;)]

> FWHK, csync2 doesn't fit very well in a multi-master environment.

?! reading the docs make me feeling the opposites...


> However, you may work around the problems you may experience in changing the
> csync2 config file by replicating it first, in a dedicated session.

Very good hint. Indeed the right solution.


> When files are the same in both the master and the slaves, you should see a
> message like:
> 	Updating /some/dir/some.file on some.slave ...
> 		File is already up to date on peer.
> When files differ in content, then csync2 updates it unless you're running
> the first sync session OR the slave file got changed with respect to the
> last sync session. In that case you get a synchronization conflict.

It is exactly the clue. In my test i had two dir already syncronized
(by hand via scp), so the files differ only by timestamps and
permissions.
Unison explicitly identify the file as the same, and propagate only
timestamps, owner and permissions (if configured to do so).

With csync2 i had the feeling that files was tranferred, not files
attributes.
My next test will do agains my 'shareware and drivers' share, that
weight some gigabytes...


> Again, csync2 is not well suited in a multi-master environments: you may
> expect a conflict if somebody changes the content of a file in a slave.

Life is full of conflicts: the trouble are not conflicts, but the facts
that unison have a really unparsable output/log, so it is very hard to
build 'something' that can, automatically/programmatically, solve the
conflict.


> You have the option to use the -I flag in order to resync the whole set of
> slaves after adding a new one.

OK. Now is clear.


[ACL migrating scripts]
> > Someone have done some sort of hacks like this?
> Not to my knowledge.

Would be useful, to implement reilably, to have in script execution
some more expansion variable, like remote server, remote file, ...


> Marco, apart your ACLs whish, which I think may eventually fit into csync2,

If i can add one more, a logging mechanism, that could be even a simple
'log to file', insted of the simple '>'.


> I have the feeling you're using the wrong tool.
> csync2 is designed to complete syncs consuming very low bandwidth, but it
> doesn't work well when changes are often expected in slaves.
> Since you're instead working on a LAN and have probably plenty of bandwidth,
> I would look for something else. Samba has recently improved DFS support, in
> example. Or you could use something like GFS. 
> Basically, you want changes to a file in a server to be immediately
> propagated to the other servers, right?

Nono, i'm working on a VPN network ADSL-based so the bandwidth is on
Kbit/s, not Mbit/s...
Simply i've some users that work one day in one office, one other in
the other and the night suffice to sync their data. ;-)

Anyway for now i'm testing it against configuration file (as just
sayed), i will test against my semi-public, semi-static repository 
of drivers and tools, and i hope i will use for the rest.


Many thanks.

-- 
dott. Marco Gaiarin				    GNUPG Key ID: 240A3D66
  Associazione ``La Nostra Famiglia''                http://www.sv.lnf.it/
  Polo FVG  -  Via della Bontà, 7 - 33078  -  San Vito al Tagliamento (PN)
  marco.gaiarin(at)sv.lnf.it	  tel +39-0434-842711  fax +39-0434-842797

		Dona il 5 PER MILLE a LA NOSTRA FAMIGLIA!
	   http://www.lanostrafamiglia.it/chi_siamo/5xmille.php
	(cf 00307430132, categoria ONLUS oppure RICERCA SANITARIA)