Cannot Unload tape on HP-UX (device busy)

We have DP5.5
The cell server runs on W2k3. It has the following patches:

Patch level Patch description
===========================================
DPWIN_00249 Core Component
DPWIN_00249 Core of Integrations component
DPWIN_00208 Cell Manager Component
DPWIN_00231 Disk Agent
DPWIN_00232 Media Agent
DPWIN_00191 User Interface
DPWIN_00191 Manager-of-Managers User Interface
DPWIN_00181 MS Exchange 5.x Integration
DPWIN_00181 MS Exchange 2000/2003 Integration
DPWIN_00209 MS SQL 7.0/2000 Integration
DPWIN_00118 Oracle Integration
Number of patches found: 11.

On clients running HP-UX we have the following patches:
Patch level Patch description
===========================================
PHSS_34497/PHSS_34498/DPSOL_00207 Core Component
PHSS_34499/PHSS_34500/DPSOL_00208 Disk Agent
PHSS_34142/PHSS_34143/DPSOL_00184 Media Agent
Number of patches found: 3.

We start having some problems on HP-UX backup since last week. Whe've experienced 3 problems:
2 - When a backup starts the bma is launched and is imediately aborted with this message:

[Normal] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:01:00
STARTING Media Agent "Drive 4"

[Normal] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:01:14
By: UMA@omnilx.corp.auchan.pt@scsi2:0:3:0
Loading medium from slot 2 to device /dev/rmt/36mn

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/oracle_client]" Time: 05-12-2006 17:01:57
STARTING Disk Agent for sxarh01:/product/oracle_client "sxarh01.corp.auchan.pt [/product/oracle_client]".

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/mqseries/data]" Time: 05-12-2006 17:02:02
STARTING Disk Agent for sxarh01:/product/mqseries/data "sxarh01.corp.auchan.pt [/product/mqseries/data]".

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/oracle_client]" Time: 05-12-2006 17:01:59
Filesystem Statistics:

Directories ........ 247
Regular files ...... 0
Symbolic links ..... 0
SYSV FIFOs ......... 0
BSD sockets ........ 0
Block Devices ...... 0
Character Devices .. 0
Unknown Objects .... 0
----------------------------
Objects Total ...... 247
Kbytes Total ...... 1209

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/mqseries/data]" Time: 05-12-2006 17:02:02
Filesystem Statistics:

Directories ........ 150
Regular files ...... 2
Symbolic links ..... 0
SYSV FIFOs ......... 0
BSD sockets ........ 0
Block Devices ...... 0
Character Devices .. 0
Unknown Objects .... 0
----------------------------
Objects Total ...... 152
Kbytes Total ...... 274

[Major] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:02:03
[90:180] Cannot rewind medium. ([16] Device busy)

[Critical] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/oracle_client]" Time: 05-12-2006 17:02:03
Unexpected close reading NET message => aborting.

[Critical] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/mqseries/data]" Time: 05-12-2006 17:02:03
Unexpected close reading NET message => aborting.

[Critical] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/oracle_client]" Time: 05-12-2006 17:02:03
Connection to Media Agent broken => aborting.

[Critical] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/mqseries/data]" Time: 05-12-2006 17:02:03
Connection to Media Agent broken => aborting.

[Major] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:02:03
[90:180] Cannot rewind medium. ([16] Device busy)

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/oracle_client]" Time: 05-12-2006 17:02:03
ABORTED Disk Agent for sxarh01:/product/oracle_client "sxarh01.corp.auchan.pt [/product/oracle_client]".

[Normal] From: VBDA@sxarh01 "sxarh01.corp.auchan.pt [/product/mqseries/data]" Time: 05-12-2006 17:02:03
ABORTED Disk Agent for sxarh01:/product/mqseries/data "sxarh01.corp.auchan.pt [/product/mqseries/data]".

[Major] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:02:03
[90:135] Cannot eject medium. ([13] Permission denied)

[Major] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:02:03
[90:64] By: UMA@omnilx.corp.auchan.pt@scsi2:0:3:0
Cannot unload exchanger medium (Details unknown.)

[Critical] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:03:44
IPC Cannot Allocate Shared Memory Segment
[Critical] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:05:25
IPC Cannot Allocate Shared Memory Segment
[Normal] From: BMA@sxarh01 "Drive 4" Time: 05-12-2006 17:05:26
ABORTED Media Agent "Drive 4"

[Critical] From: BSM@omnilx.corp.auchan.pt "sxarh01" Time: 05-12-2006 17:09:39
None of the Disk Agents completed successfully.
Session has failed.

[Normal] From: BSM@omnilx.corp.auchan.pt "sxarh01" Time: 05-12-2006 17:09:39

Backup Statistics:

Session Queuing Time (hours) 0,00
----------------------------------------
Completed Disk Agents ........ 0
Failed Disk Agents ........... 21
Aborted Disk Agents .......... 0
----------------------------------------
Disk Agents Total ........... 21
========================================
Completed Media Agents ....... 0
Failed Media Agents .......... 1
Aborted Media Agents ......... 0
----------------------------------------
Media Agents Total .......... 1
========================================
Mbytes Total ................. 0 MB
Used Media Total ............. 0
Disk Agent Errors Total ...... 4

On this case i have ran the backup on debug mode and on the log i find this and when it tries do rewind the tape there is the following errors:

[ 31] ===== START [37] MESSAGE FROM CON
[ 31] |1
[ 31] |0
[ 31] ===== END
[ 31] MaReactToEvent: from CON: received MSG_SEEK (LO=5, LI=3)
[ 31] MaEventLoop: state returned: 12
[ 11] MaStateProcess: BMA maState: STATE_SEEK
[ 31] MaStateProcess: from CON got: MSG_SEEK(,)
[ 61] SeqSeek: (whence:0, seg:0, blk:0)-START: curPos=0:0
[ 61] SeqOp: op=5, num=1
[ 61] SeqOp: II: op=5, num=1
[ -2] 2006-12-05 17:01:52 ("ma/dev/devseq.c /main/dp55/dp55_fix/6":533) A.05.50 bPHSS_34142_PHSS_34143_DPSOL_00184
[ -2] SeqOp: (/dev/rmt/36mn): ioctl(MTIOCTOP, mt_op=5, mt_count=1) fails: {16}

[ 4] 2006-12-05 17:01:52 ("ma/dev/devseq.c /main/dp55/dp55_fix/6":550) A.05.50 bPHSS_34142_PHSS_34143_DPSOL_00184
[ 4] [MARK 2] in SeqOp
Code is:1026 SystemErr: [16] Device busy



---------------------------------------------------

2 - The backup starts fine but when it reches the end the BMA tries to eject the drive and gives this error:

[Major] From: BMA@machine "Drive 2" Time: 05-12-2006 16:04:57
[90:135] Cannot eject medium. ([5] I/O error)

[Normal] From: BMA@machine "Drive 2" Time: 05-12-2006 16:04:57
/dev/rmt/7mn
Tape Alert [10]: You cannot eject the cartridge because the tape drive
is in use. Wait until the operation is complete before ejecting the cartridge.

[Major] From: BMA@machine "Drive 2" Time: 05-12-2006 16:04:57
[90:64] By: UMA@machine.domain@scsi2:0:3:0
Cannot unload exchanger medium (Details unknown.)










==============================================

After this errors occur the drive stays on a state that only "power off/on", release the drive for other machines. i've tried mt commands to offline the tape but it gives me just "I/O Error".

We've ran the hp_ltt tool and on the healthcheck of the drive, he've saw that the drive is locked by an application even after the end of the backup.

It seems that it's the drive is stuck on some BMA operations.
There are sometimes that these problems doesn't happen, which is kind of a strange behaviour.
On our Solution we have Windows, AIX and HP-UX systems, and this problem only occurs on HP-UX, even on different types of machines.

We've contacted the HP Support and we're waiting for a callback but since these problem locks the drives, it's urgent to solve this issue.

Can you give some kind of help on this please?