Hi, Thanks for your response. Please see my questions line
On Thu, Apr 15, 2010 at 10:37:29AM -0700, Paul Yeatman wrote:
> Hi!
>
> On Thu, 2010-04-15 at 10:18 +0100, Mark Adams wrote:
> > Hi All,
> >
> > Debian lenny, Amanda 2.6.1p1-2
> >
> > I'm backing up a single 1.9T xfs filesystem to an LTO4 drive with
> > changer. The backup seems to complete correctly and recovering files
> > that are on the first tape complete correctly.
> >
> > However, when trying to retrieve and files that are on the 2nd tape, the
> > tape is loaded correctly by the changer (after first spooling through
> > the first tape, as a side note is there a way to skip this?) then after
> > some time the following error is shown
>
> You would not be able to skip the first tape if this is where the backup
> image begins
Ok. It doesn't help if it is in 80G chunks?
>
> > tar: Error exit delayed from previous errors
> > Extractor child exited with status 2
> >
> > How can I find out what the "previous errors" were?
>
> You need to look in the debug log for the backup application you are
> using on the client in /var/log/amanda/client/<config>/.
I don't have any amgtar logs, just the amrecover logs - which have the
security stream_stream_close message I noted below. Is there some
additional logging I can add?
>
> For amgtar, it will be amgtar.<datestamp>.debug
>
> Cheers,
> Paul
>
> >
> > In the amrecover log I also see
> >
> > 1271275442.543746: amrecover: security_stream_close(0xd89180)
> > 1271275520.583556: amrecover: security_stream_close(0xd99f30)
> >
> > My tape and dump config is as follows
> >
> > define tapetype LTO4 {
> > comment "HP LTO4"
> > length 804191104 kbytes
> > filemark 0 kbytes
> > speed 90511 kps
> > blocksize 32 kbytes
> > }
> >
> > define dumptype upbackup {
> > program "GNUTAR"
> > compress server custom
> > server_custom_compress "/usr/bin/pigz"
> > tape_splitsize 80Gb
> > split_diskbuffer "/tapehold/"
> > fallback_splitsize 10Gb
> > comment "UPBACKUP"
> > index
> > priority high
> > auth "bsd"
> > }
> >
> > Help would be very much appreciated!
> >
> > Cheers,
> > Mark
