Hi All
We have a situation that I am struggling to resolve, even with
the aid of symantec.
All NBU 6.5.3.1
Solaris 10 cluster Master
Solaris 10 media server
HPUX 11.00 media server
SUN VTL Plus 2 with NDMP 4 functionality
L700 with 4 stk T10k drives with encryption – mounts/dismounts
via ACSLS
The basis of what we are trying to achieve( which is working on
the most part) is by writing the primary copy to a vtl volume and then duplicating
a copy to physical with the VTL (as the NDMP host) being the data mover for the
duplication. This is all working fine and indeed the vtl is moving the data
during the duplication process, in turn alleviating the responsibility from the
NBU media server, however the issue we have is that most of the time not all 4
backend drives are being utilised for no apparent reason. Most of the time ¾ are
being used, and sometimes 2/4.
Each media server is presented a virtual library along with a
number of vtl drives for the primary copy of a SLP.
Each media server is then configured via NDMP for the use of the
T10k physical drives.
Not to complicate the matter with the configuration as the main
issue is not utilising 4/4 backend drives. All the main suspects have been
checked, i.e. max drives on the stu, max jobs per client etc etc. think this is
proved as there are occasions where 4/4 do get utilised
The backend T10k’s are shared between the media servers,
i.e. NDMP SSO , vmoprcmd extract below can see only ¾ being used. Additionally
vmdareq below that
NDMP_Dr00-01 Yes Yes 200594 200594
Yes hcart3
$hpux_server /dev/scsi/sequential/c2t0d0
($vtl_node_with_ndmp4) ACS
$solaris_server /dev/scsi/sequential/c2t0d0 ($vtl_node_with_ndmp4)
ACTIVE
NDMP_Dr00-02 Yes Yes 200589 200589
Yes hcart3
$hpux_server /dev/scsi/sequential/c3t0d0
($vtl_node_with_ndmp4) ACS
$solaris_server /dev/scsi/sequential/c3t0d0 ($vtl_node_with_ndmp4)
ACTIVE
NDMP_Dr00-03 Yes Yes 200590 200590
Yes hcart3
$hpux_server /dev/scsi/sequential/c4t0d0
($vtl_node_with_ndmp4) ACS
$solaris_server /dev/scsi/sequential/c4t0d0 ($vtl_node_with_ndmp4)
ACTIVE
NDMP_Dr00-04 No No
No hcart3
$hpux_server /dev/scsi/sequential/c2t1d0 ($vtl_node_with_ndmp4)
ACS
$solaris_server /dev/scsi/sequential/c2t1d0 ($vtl_node_with_ndmp4)
ACS
NDMP_Dr00-01 - RESERVED on Wed May 20 11:18:10 2009
$hpux_server UP
$solaris_server RESERVED UP
NDMP_Dr00-02 - RESERVED on Wed May 20 11:17:57 2009
$hpux_server UP
$solaris_server RESERVED UP
NDMP_Dr00-03 - RESERVED on Wed May 20 11:20:27 2009
$hpux_server UP
$solaris_server RESERVED UP
NDMP_Dr00-04 - AVAILABLE
$hpux_server UP
$solaris_server UP
From my assumptions I think there is still a reserve of some
sort being held on the drive that isn’t being used, attempted to scour a
“nbrb –dump” for this particular drive to no avail.
So if anyone has experienced anything similar or indeed can
share some advice etc I would be most grateful.
Regards
David Clooney