RFE: using rsync as a backup tool (preserve access time & compres
s destination files) ?
Gilles-Eric.Descamps at SiliconAccess.com
Wed Oct 30 00:57:00 EST 2002
I know those questions have been asked before
but that was more than an year ago.
I'm hoping the situation has evolved now.
I'd like to use rsync as a backup tool
to move around some data.
I often have to move hundreds of GB,
and that takes some time. I'd like to use rsync
so that if something happens, I can restart
the migration without loosing what has already been transfered.
But I noticed that every time I use rsync to backup
some data, it "corrupts"/updates the access time of the source file.
All the backup tools I've previously used preserve the access time
(otherwise, you'll loose your access time if you backup everyday).
Gnu tar has a "--atime-preserve".
I use both the access time (last read) & modification time (last write)
to identify which files to remove from the fast file server.
(I've written a tiny perl script to identify which files are "old").
I wish there was an rsync option to preserve access times (sce & dst).
I would also be delighted if there was another option that would
allow to compress/gzip/bzip files on the destination.
I believe that by having almost same filename -> filename.gz,
and having all times (A, C, M) identical, would be sufficient for a match ?
Why compressing files ? it's the best trade-off between ease of use
and compression. having no compression at all for old files leads to
waste of space (even on a cheap slow file server).
Compressing everyfile into a big tarfile makes it difficult to browse
a huge archive (I don't know of a "tar shell").
Keeping the directories uncompressed allows for browsing the data easily.
Thanks in advance,
-------------- next part --------------
HTML attachment scrubbed and removed
More information about the rsync