Blah. Sorry for the noise.
I had forgotten that the LTO6 was still getting assigned /dev/nst0. So,the LTO7 drives are /dev/nst1
and /dev/nst2 (not 0 and 1). I don't know what was transpiring with the LTO7 configuration in amanda
trying to read something from the LTO6 drive, but when I changed to recovering from the correct
drive, it worked as expected. And I was worried about Amanda keeping its brain straight. ;-)
On 1/27/20 2:44 PM, Chris Hoogendyk wrote:
I'm running into some confusion running amrecover with Amanda version 3.5.1.
I have a slightly complex amanda.conf with storage definitions that I originally set up when I was
running two tape libraries, one with an LTO6 drive and one with two LTO7 drives. I've
decommissioned the LTO6 now, but I still have the storage definition that is used for the LTO7.
The backups seem to be running just fine and sometimes will overlap with parallel instances. I
don't quite understand how Amanda keeps its brain straight when this is happening, and maybe it
doesn't entirely.
When I use the amtape command, I will typically do something like "amtape -o storage=research
daily slot 24". The two drives show up as /dev/nst0 and /dev/nst1. I have amanda-client.conf set
to /dev/nst0, knowing that I can change that on the fly to /dev/nst1.
With amrecover, apparently there is no facility to indicate the storage. I've tried and I've
looked at the man page. When I simply do the amrecover without that specified, I keep getting
amrecover> extract
Extracting files using tape drive tape:/dev/nst0 on host marlin.
The following tapes are needed: Bio-Research-027
Extracting files using tape drive tape:/dev/nst0 on host marlin.
Load tape Bio-Research-027 now
Continue [?/Y/n/s/d]? y
Error reading Amanda header: block size too small
Load tape Bio-Research-027 now
Continue [?/Y/n/d]?
I'm not sure where to go with this. Unfortunately, I have to get it ironed out pretty quickly. I'm
retiring at the end of this week, and I need to provide documentation on how to do everything.
This works on my other Amanda servers, but this server is more complicated.
I am copying the amanda.conf below. The tail of the amrecover log is at the
very end.
--
---------------
Chris Hoogendyk
-
O__ ---- Systems Administrator
c/ /'_ --- Biology & Geosciences Departments
(*) \(*) -- 315 Morrill Science Center
~~~~~~~~~~ - University of Massachusetts, Amherst
<[email protected]>
---------------
Erdös 4
