Date: Sat, 17 Oct 2009 19:21:31 -0400 From: Jerry McAllister <jerrymc@msu.edu> To: Tim Judd <tajudd@gmail.com> Cc: Jerry McAllister <jerrymc@msu.edu>, freebsd-questions@freebsd.org Subject: Re: small question about tape-based dumps Message-ID: <20091017232131.GB66093@gizmo.acns.msu.edu> In-Reply-To: <ade45ae90910161943o7695d436wed8b8a69d92ef994@mail.gmail.com> References: <1255727601.4640.4.camel@x1-6-00-11-09-00-e4-00.search.b.superkabel.de> <20091016213732.GA61433@gizmo.acns.msu.edu> <ade45ae90910161943o7695d436wed8b8a69d92ef994@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Oct 16, 2009 at 08:43:26PM -0600, Tim Judd wrote: > Replies inline > > On 10/16/09, Jerry McAllister <jerrymc@msu.edu> wrote: > > On Fri, Oct 16, 2009 at 11:13:21PM +0200, Stevan Tiefert wrote: > > > >> Hello list, > >> > >> one example: If I have three partitions and I want to backup every day > >> these partitions, will I need 21 tapes? > >> > >> I ask because it seems it is not possible to place more than one dump on > >> one tape, isn't it? > > > > You can easily put more than one dump on a tape if there is > > room enough for them. Check out the mt(1) command. > > > > Something like mt fsf 1 will skip over the first dump file > > so you can write the second. mt fsf 2 will skip over two files, etc. > > That is dump files, not files within the dump. Each dump of a > > filesystem is one file. > > > > If you need to restore, it is just the same. The first dump is > > the first file. The second dump is reached by skipping 1 file > > with the mt command, etc. > > > > I actually rewind and skip between each dump of multiples made > > to the same tape. I also use the no-rewind device for the tape. > > > > So first dump is: dump 0af /dev/nsa0 / > > I understand that this creates a dumpfile on nsa0, and as I understand > tapes (which may be wrong, which I ask for clarification here).. To > mark a end-of-file to be able to fast-forward/rewind, why can't you > use: > mt -f /dev/nsa0 weof > > It's description in mt(1) says it writes the end-of-file mark at > current position You do not need to. dump alrady writes that when it finishes each time. If you to that, you will get a second one at that location. You do not need to do the rewind and mt fsf between each dump. I just do it to make it very clear to myself in my scripts what I am expecting and that I am doing it right. ////jerry > > > For second dump: mt -f /dev/nsa0 rewind > > mt -f /dev/nsa0 fsf 1 > > dump 0af /dev/nsa0 /usr > > So if we use weof, would the 2nd dump then be: > dump 0af /dev/nsa0 /usr > mt -f /dev/nsa0 weof > > > third mt -f /dev/nsa0 rewind > > mt -f /dev/nsa0 fsf 2 > > dump 0af /dev/nsa0 /var > > And 3rd: > dump 0af /dev/nsa0 /var > mt -f /dev/nsa0 weof > > > etc. > > > > when all done mt -f /dev/nsa0 rewind > > mt -f /dev/nsa0 offline > > And I've never used offline, guess I'll start now. > > > I have this all in a script that also writes an index file > > as the first file on the tape. > > > > Of course if you are doing a change dump the dump command is > > going to look more like: > > > > dump 1af /dev/nsa0 > > etc. > > > > ////jerry > > > >> > >> With regards > >> Stevan Tiefert > > > > Thanks for any input! > --TJ
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091017232131.GB66093>