Greetings.
Amanda has been running great at our site, backing up 28 file systems for 6
hosts
for the past year and a half. However, there are 2 questions I have about
amanda.
1) We run Veritas Cluster Server, which has the ability to move
a file system along with its associated IP address. When we first tried to use
amanda, amcheck reported errors eprod and prod1 were both names pointing to
the same "eprod" host - and you are not allowed to run amandad twice.
Our kluge at that time was to write a translation script that would
take a disklist and run hostname on each of the hosts in disklist,
and write a new disklist with actual hostnames.
This makes the assumption that a failover won't happen during the backup.
The better thing would be to allow amanda to back up
prod1:/dg/prod1 even though the real hostname of prod1 is "eprod".
I am curious if anyone has encountered this, looked into it?
2) During Disaster Recovery Exercise, it would be really great to
use amrecover to read the tape. But of course, a DR exercise assumes
that the backup host is gone. I found the script showing how to
list tape contents with "dd if=/dev/rmt/0n bs=32k count=1 | head -1",
(Until I discovered the dd trick, I was afraid my nightly amverify
would have to be printed and accompany the tape....fortunately not)
If I could just restore my index/tape host first, it would be better.
The problem is, the order of the dumps is not constant.
Is there a way to force it to put the index host on the tape first?
Thanks
Seth
|