Hi folks,

I'm trying out Flashbackup from Veritas (Netbackup Datacenter option). It works 
well for the major part of our fileserver spare the partitions that are located 
on an A1000 unit from Sun. We are running Solaris 8, and the errors we are 
getting look like the following (very verbose logging will generate a log that 
100mb in size...):

13:00:48.697 [3085] <4> bpfsmap: INF - print_imgmapent: ino=10 type=2 
size=[10794,0]
13:00:48.697 [3085] <16> bpfsmap: ERR - ufs_put_meta: shadow inumber for file 
11 is 0
13:00:48.697 [3085] <16> bpfsmap: ERR - put_tapefil: OP_PUT_META returned error 
-1
13:00:48.698 [3085] <16> bpfsmap: ERR - put_tapefil: error writing inumber 11 
file 11 to tape - backup of this file aborted

and on and on it goes for every single inode. Some people reported that the 
"rdaemon" and "xntp" processes could conflict and put the array in "locked" 
mode. However, killing them doesn't solve it.

Here's how my flashbackup job is configured:

CACHE=/dev/md/rdsk/d5
/dev/rdsk/c1t5d0s6

The slice we are trying to backup is mounted using /dev/rdsk/c1t5d0s6 (but 
that's a non-issue since flashbackup operates with mounted slices). If someone 
can help me here, that would be extremely helpful (and beneficial for Veritas 
since this product is much faster for 2 millions of small files).

Thanks,

-- 
Mat Nantel              
Ecopia BioSciences      
IT Security Administrator
(514) 336-2724 x434