Re: [rdiff-backup-users] ACLS

[Ben Escoto]

>>>>> "GF" == Greg Freemyer <address@hidden>
>>>>> wrote the following on Tue, 11 Feb 2003 16:02:42 -0500

  GF> I've just discovered your project.  In your opinion, is it
  GF> production ready, or does it have limitations like "no files
  GF> over 2 GB".

Well I don't know about "production ready", but it seems to work for a
number of people (including me).  Right now I am working on the error
policy, which has changed between various versions.  I am pretty sure
now that version 0.11.2 will have session-level atomicity.

(Note to people on list who discussed this earlier:  I did write some
journaling code, but in the process realized that it will probably be
simpler to make sure all the changes are written in a predictable
order.  The session after a crash will then do the equivalent of an
fsck and correct all the errors.  Having the metadata makes this much
easier.)

    About the 2GB limit, I was recently informed of a new problem.
See:

http://sourceforge.net/tracker/?group_id=5470&atid=105470&func=detail&aid=683061

The upshot is python's gzip module cannot write compressed files
longer than 2 (4?) GB.  So files greater than 2GB are ok as long as
rdiff-backup never has to write a large compressed volume.  (This
would be the case if, for instance, the large file was very
compressable or didn't change much between sessions.)

One workaround is to use --no-compression[-regexp] to make sure your
large files don't get compressed.

  GF> Are you handling ACLs?  I ask because I want to backup a Samba
  GF> Server running on an ACL enable FS.  (i.e. XFS)

  GF> If your not yet handling ACLs, is in the plans?

I don't see any reason not to handle ACLs, but my system doesn't have
them at the moment.  Maybe in 0.13.0.  


-- 
Ben Escoto

pgpXL65lTiu6V.pgp
Description: PGP signature