On Jun 7, 2005, at 12:54 PM, Tom Duerbusch wrote:
I'm reading up on Amanda, well, currently some old documentation
(SG246299 Linux on IBM eserver zSeries and S/390 ISP/ASP Solutions).
Using Amanda in a traditional network world, Amanda would be
running on
a server with tape drives and the clients will be running on other
images and the backups are taken over the network. Sure, saves on
tape
drives.
But on a mainframe, each image can own the tape drives.
I'm wondering (pros and cons) if anyone has made every Linux image, an
Amanda server and just back up its self. No network envolved. It
seems
like it would be more efficient, but then....I think there would be a
lot of unknowns of how things work when you actually need something
restored (or your image rebuilt).
Or maybe Amanda just won't run this way.
It'd run fine, except....
a) how do you ensure that the tape drive is actually attached to the
right guest?
b) how do you manage your tape catalogue? Amanda would want one tape
set per instance in this case. How do you get the RIGHT tape loaded
into the right drive when Amanda needs it?
We've actually solved a lot of this problem using Bacula rather than
Amanda, by building a little tape server app that talks to VM and
asks it to do the tape manipulation.
Under Amanda, you could do it with HSM and NFS-exported managed
storage, but you still have the problem of needing to dedicate fake-
tape space to each guest.
Basically, it's not a backup problem, it's a tape management problem,
in your scenario.
Adam
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
