Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 25 Mar 2009 17:19:03 +0000
From:      Hywel Mallett <hywel@hmallett.co.uk>
To:        freebsd-fs@freebsd.org
Subject:   Re: dump | restore fails: unknown tape header type 1853384566
Message-ID:  <20090325171903.x4eqs9ceo8w00c04@www.hmallett.co.uk>
In-Reply-To: <49CA57BB.2070409@andric.com>
References:  <49C83673.3000604@aldan.algebra.com> <200903241537.36515.doconnor@gsoft.com.au> <49C87E0D.5090501@aldan.algebra.com> <49CA57BB.2070409@andric.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Quoting Dimitry Andric <dimitry@andric.com>:

> On 2009-03-24 07:30, Mikhail T. wrote:
>>     dump a0f  - /old | restore -rf -
>>     [...]
>>       DUMP: 17.25% done, finished in 3:27 at Tue Mar 24 05:42:00 2009
>>       DUMP: 20.36% done, finished in 3:09 at Tue Mar 24 05:28:13 2009
>>       DUMP: 23.83% done, finished in 2:50 at Tue Mar 24 05:14:32 2009
>>     unknown tape header type -621260722
>>     abort? [yn]
>>
>> Looks like a junk value somewhere... Unitialized variable or some such.

Looking at the restore code (tape.c, in findinode), restore is =20
expecting a header type in the range 1-6, so the header type =20
-621260722 is way out. Assuming that findinode is being passed the =20
correct variable, it would indicate that dump is writing the header =20
(or at least the header type incorrectly). I can't work out where this =20
header is getting written though. It looks like plenty of data gets =20
written into a header (such as inode, magic number, checksum). I =20
wonder if one of these values is overflowing and overwriting the =20
header type?

>
> Maybe the dump output gets corrupted in some way?  (E.g. faulty RAM, or
> disk?)  If you are dumping a live filesystem, could it possibly help to
> add the -L option?

It might be worth fscking the original volume (though I suspect the OP =20
has done this already), and also passing the -D option to restore, as =20
restore will then try and continue, rather than abort on getting the =20
invalid header type. Fixing the root cause would be better, but that =20
might be a workaround for now.




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090325171903.x4eqs9ceo8w00c04>