This is very bad.  "magic" errors usually indicate physical corruption in
the data.
It can be due to a physical I/O error on the media, or software, or
firmware.

I suggest you open a severity 1 PMR with IBM right away, because your
backups may be creating MORE unreadable data all the time.
IBM can tell you whether an AUDIT of all your storage is needed, or whether
there is something else you should do.

Here is a similar hit from the IBM data base:
http://www-01.ibm.com/support/docview.wss?uid=swg21406169


On Fri, Feb 19, 2010 at 8:44 AM, Pawlos Gizaw <
Pawlos.Gizaw AT sanofi-aventis DOT com> wrote:

> Has any one seen the below error?  I have TMS 5.5.4 running on AIX 6.1
> and no issue on all backups except backups pointed to data domain
> storage pool. Recently when DBA  try to restore rman backup getting
> message  ANS0322E (RC40) no text available for this return code. When I
> check the TSM server side  see the below message. After the first
> incident datadomain/emc swapped out the hardware but we are still
> getting same error.
>
>
>
>
>
> ANR9999D_2590516390 NtpValidateComBlockHdr(pvrntp.c:5665)
>                          Thread<84518>: Invalid block header read from
> NTP drive
>                          DRIVE17 (/dev/rmt47).(magic=5A4D4E50, ver=5,
> Hdr
>                          blk=241976 <expected 241977>, dbytes=262096
> <262096>)
>                          (SESSION: 67666, PROCESS: 232)
>  ANR9999D_2590516390 NtpValidateComBlockHdr(pvrntp.c:5665)
>                          Thread<84518>: Invalid block header read from
> NTP drive
>                          DRIVE17 (/dev/rmt47).(magic=5A4D4E50, ver=5,
> Hdr
>                          blk=241976 <expected 241977>, dbytes=262096
> <262096>)
>                          (SESSION: 67666, PROCESS: 232)
>  ANR0541W Retrieve or restore failed for session  - error on input
> storage
>
>
>
> Thanks
> Pawlos
>