First thing, check the dump logs to make sure it dumped the root file
system.
Second, check the amanda indexes to make sure the root filesystem is there
Third, try to restore it by using amrestore, instead of amrecover
Fourth, try to restore it by using tar, or dump, directly from the tape,
without using amanda

Fifth, if all the above fails, try doing a manual dump and restore on that
filesystem

-----Original Message-----
From: Dan Debertin [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, May 08, 2002 12:02 PM
To: [EMAIL PROTECTED]
Subject: Root directory is not on tape?


I'm getting the following odd behavior when doing a restore:

amrecover> extract

Extracting files using tape drive /dev/nrst0 on host flame.nodewarrior.org.
The following tapes are needed: DLT1

Restoring files into directory /tmp
Continue? [Y/n]: y

Load tape DLT1 now
Continue? [Y/n]: y
Root directory is not on tape
abort? [yn] y

The amanda server is a NetBSD-1.5.3 machine, and the dump is ufsdump
from a Solaris 2.5.1 machine. There are two other similar slices on
this dump from the same machine (/var and /) that restore just fine,
so I'm boggled as to why this one fails. It is a rather large slice --
about 11GB (/usr/local), where the other two are roughly 500MB
each. The filesystem was completely quiescent -- I took one mirror of
a RAID0+1 array offline and backed that up, so no writes took place
while the dump ran.

It appears that my "restore" is what's generating this message, not
amanda itself. But I'd appreciate some insight into what amanda did
(if anything) to create a dump that restore can't read.

Dan
--
[EMAIL PROTECTED]
www.nodewarrior.org
ignorami: n: 
The art of folding problem users into representational shapes.

Reply via email to