You are on page 1of 8

AMS2000

MICROCODE VERSION 0885/A


RELEASED 01/14/2010

Newly supported features and functions for Version 0885/A

New features and functions


Contents 1) New HDD support (2TB SATA, 200GB SSD, 600GB 15K SAS)
DF-F800-AVE2K 3.5" HDD 1968.52GB( 7,200) -SATA 3276139-D
DF-F800-AVE2KX (Dense 3.5" HDD 1968.52GB (7,200) - SATA 3282101-D
RKAKX)
DF-F800-AKS200 3.5" SSD 195.82GB - SSD 3282195-B
DF-F800-AKH600 3.5" HDD 575.30GB (15,000) - SAS 3276138-D

Improvements for Version 0885/A

1 9 Improvements
Contents 1) Improvement of failure search accuracy against the illegal access from one controller to the other
causing “[DCTL]The bridge error was detected [external access]”
2) Improvement to check process of the firmware update
3) “SATA Write & Compare” check (or SATA RAW) can be disabled or enabled by HSNM2
4) Improvement of failure search accuracy of backend diagnosis. Add the SAS address consistency check
between SAS expander firmware, controller and ENC
5) Improvement of DP Pool recovery speed
6) Improvement of capturing dump data for the iSCSI interface. The dump data of iSCSI interface can be
captured in a week without PS OFF/ON or firmware update
7) DP page information is saved to drives at firmware update
8) Improvement of response time when Queue full occurs frequently on FC interface
9) Improvement of failure search accuracy of SAS controller firm

The change of contents for Version 0885/A

1 The subsystem may go down due to write failures of a DMLU


Phenomena The subsystem may go down due to write failures of a DMLU accompanied by a controller
failure during shutting down or rebooting the subsystem.
Severity High
Conditions of Occurrence The problem may occur when all following conditions are met.
(1)When the subsystem is being shut down or when the subsystem is being rebooted.
(2)A controller failure occurs while write operations to a DMLU are being implemented.
Cause When a controller failure occurs during shutting down or rebooting the subsystem, even if
the controller failure does not affect write operations to a DMLU actually, the write
operations are considered a failure and the subsystem goes down.
Affected Products/Version DF800H/M/S, All version
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

HDS Confidential 1 of 8
2 HDD failure caused by LA/LRC errors
Phenomena In case that DF800 receives the read/write command from hosts, blockading multiple HDU
may occur with following messages.
IY0F00 LA error (errcode:LUN) (Unit-x, HDU-y)
IY0G00 LA error (errcode:LBA) (Unit-x, HDU-y)
IY0H00 LRC error (Unit-x, HDU-y)
x: Unit number, y: HDU number

With DF800/SA8x0 managed by Hitachi Device Manager (HDVM) and this phenomenon
occurs on this DF800/SA8x0, following alert message will be shown in HDVM.
Serious error reported for Disk Drive #n.
In addition, following log message will be shown in Event Log (Windows) and/or System
Log (Other OS) where HDVM has been installed.
KAID11010-E Serious error reported for Disk Drive #n.
Severity High
Conditions of Occurrence The problem may occur when all following conditions are met.
1. If DF800 the quick format command is ordered to a LU of more than 48TB (This applies
to multiple LU of more than 48TB specified, too.)
2. If SA8x0 the quick format command is ordered to a LU of more than 4.8TB (This
applies to multiple LU of more than 4.8TB specified, too.)
3. Executing the quick format to the first 64GB area where 48TB (DF800H) or 4.8TB
(SA8x0) is exceeded.
4. SNM2/Hicommand is available with the array.
Cause When the subsystem receives a sense command through the LAN while formatting LU in
the quick mode, an error relating to the sense command occurs and some spaces become
unformatted.(transfer information that is supposed to be held in the local memory is
mistakenly held in the quick format managing table in the cache memory. )As a result,
when read/write come to the unformatted spaces, several HDD failures are falsely detected
by LA/LRC errors.
Affected Products/Version DF800H, V6.2/A(0862/A) or V6.5/A to V6.5/H(0865/A to 0865/H)
SA8x0, V5.0/A to V5.0/E(1850/A to 1850/E)
Fixed Product/Version Version: x885/A or later
Category Common
Changed Part MICRO-Disk

3 Subsystem down caused by controller blockade when HDP is enabled.


Phenomena If controller is blockaded (by reason other than Power Supply failure) when HDP is
enabled, the other controller may block with a message "Microprogram error [HDP]".
Severity High
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) HDP is enabled
(2) Power Source outage (without cache volatile) or controller blockade (other than Power
Source failure) occurs
Cause The firmware manages chunk and page information of HDP as plural tables. When the
controller is blockaded or Power Source outage occurs, such tables may become
inconsistent according to the timing of the blockade/outage. Then the other controller
which is not blockaded will recover from the inconsistency. However, this recovery
process has a bug and the controller will be blockaded too. (This phenomenon does not
occur at the controller blockade caused by Power Source failure.)
Affected Products/Version DF800H/M/S, V7.0/A(0870/A) or later
Fixed Product/Version Version: 0885/A or later
Category HDP
Changed Part MICRO-Disk

HDS Confidential 2 of 8
4 System down occurred after a Power outage/Controller blockade with
COW/TCE) pairs released
Phenomena When COW Snap Shot(COW)/True Copy Extended Distance (TCE) pairs were released,
controller blockade or power outage occurred the following phenomena might be generated
(1) Subsystem down occurs.
(2) Illegal display of quantity consumed of pool and ratio consumed of pool.
(3) The Web message "the number of Pool users threshold is over" is displayed illegally.
(4) PSUE transition of pair of COW/TCE/SDR
(5) The cycle trace information of TCE is illegal.
(6) The character string of the Web message becomes illegal.
(7) With iSNS used for iSCSI interface, "Micro Program Error [ILM]” detected wrongly.
Severity High
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) COW or TCE Program Product is enabled
(2) COW or TCE pair is existing
(3) Write IO is executed
(4) Cow pair release with 1:1 or TCE pair release to which COW and cascade are not done
(5) Within 30 hours after (4), Power Source outage or subsystem down occurs
Cause Because the control information was updated illegally in cache at the error recovery
opportunity (controller brocade or Subsystem down).
Affected Products/Version DF800H/M/S, All version
Fixed Product/Version Version: 0885/A or later
Category COW/TCE
Changed Part MICRO-Disk

5 Array booting up failure in the Model upgrading process to DF800H


Phenomena In the process of the Model upgrading to DF800H, Array booting up is failed. (Can’t see
the Ready status.) One controller failure (H90320 Watch-dog time-out) is occurred and
another controller can not be accessed by LAN.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) Cache Partition Manager Program Product is enabled.
(2) Model upgrading from DF800S with 4GB/controller to DF800H with 8GB/controller.
Or, Model upgrading from DF800M with 8GB/controller to DF800H with 8GB/controller,
10GB/controller or 12GB/controller.
Cause Generally in the model upgrade, the Firmware checks if the target Array has more than
twice size from original Array and if not, rejects the operation. However, there are some
missing cases which are actually unsupported configuration in this checking logic when
4GB-DIMM is used. In this problem, access to the memory area where actually does not
exist causes the endless-loop and controller failure, so that you cannot see the Ready status.
One controller is blocked for endless-loop. Another controller can not be accessed by LAN
for endless-loop.
Affected Products/Version DF800H/M/S, V4.5/A(0845/A) or later
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

HDS Confidential 3 of 8
6 Subsystem down caused by Array booting after power cut
Phenomena Subsystem down caused by Array booting after power cut without cache volatile while
executing DP pool optimization.
Severity High
Conditions of Occurrence The problem may occur when all following conditions are met.
1. Execute DP Optimization.
2. Array booting after power cut without cache volatile or controller blockade.
Cause Logical contradiction is occurred at the timing of Array booting after power cut without
cache volatile or controller blockade during page migrating transaction in DP pool
optimization.
Affected Products/Version DF800H/M/S, V8.0/A(0880/A) or later
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

7 Plural ENC failures during Online Firmware upgrading process


Phenomena While in Online Firmware Upgrading, when one ENC is failed, then the other ENC in the
same path may be also failed. (Firmware inappropriately blocks the other ENC.)
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) ENC Firmware Upgrading process is executed on plural ENC in the same path. (Online
Firmware Upgrading or installing plural ENC at the same time.)
(2) While ENC reboot process in above (1), a certain link failure occurs.
(3) Within 15ms from above (2), internal failure transaction works.
Cause Because of inappropriate transaction for link failure in ENC, the Firmware detects the
normal ENC as failure.
Affected Products/Version DF800H/M/S, V3.2/A(0832/A) or later
SA8x0, All version
Fixed Product/Version Version: 0885/A or later
Category Backend
Changed Part MICRO-Disk

8 Controller blockade with failure in the Local Memory


Phenomena Controller blockade with failure in the Local Memory caused by Firmware logical
contradiction. One of the following messages is issued at that time.
[CSDS]The uncorrectable error on the write was detected[0xYYYYYYYY](CTL-X)
[CSDS]The uncorrectable read memory error was detected[0xYYYYYYYY](CTL-X)
The processor system bus error was detected[0xYYYYYYYY](CTL-X)
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) 1bit error (ECC correctable error) in the Local Memory occurs.
(2) 2bit error (ECC uncorrectable error) in the Local Memory occurs.
Cause Logical contradiction occurs when 1bit error (ECC correctable error) in the Local Memory
occurs, then 2bit error (ECC uncorrectable error) in the Local Memory occurs
Affected Products/Version DF800S/M/H,SA8x0 Version: All version
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

9 Controller or ENC blockade


Phenomena Controller or ENC blockade with simultaneous occurrence of processor failure and
HDS Confidential 4 of 8
backend failure.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) Temporal processor failure occurs in the controller.
(2) At the same timing of above (1), a certain Backend failure occurs in the other side of
controller.
Cause Inappropriately judging the controller or ENC as failure.
Affected Products/Version DF800H/M/S, V3.2B(0832/B) or later
Fixed Product/Version Version: 0885/A or later
Category Backend
Changed Part MICRO-Disk

10 Copy, host I/O and reserve command may error when changing LU
ownerships
Phenomena When LU ownerships are changed like during load balancing while commands are being
executed, a command error is detected, I/O operations cannot be executed or a reserve LU
cannot be set.
(a)Copy pair operations or a reserve command cannot be executed.
(b)LU become inaccessible to hosts for I/O.
When a True Copy pare is created, the status of another LU at a remote site is changed to
COPY status by mistake, so the host at the remote site becomes inaccessible for read/write.
(c)A reserve LU cannot be set.
Because another LU is reserved by mistake by Persistent Reserve OUT command, the
specified LU cannot be reserved.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
[(a)A command error occurs]
(1)A copy pair command is executed.
(2)The ownership of the specified LU is changed. (like when load balancing occurs)
(3)After the LU ownership has been changed, the LU receives a PIO related command
(other than RD/WR).
[(b)The host is inaccessible, (a)A command error occurs]
(1)A True Copy pair command is executed
(2) The ownership of the specified LU as the S-VOL by the command at the remote site is
changed. (like when load balancing occurs)
(3)After the LU ownership has been changed, the LU receives a PIO related command
(other than RD/WR).
[(c)The reserve information is wrong, (a) A command error occurs]
(1)Persistent Reserve OUT command is executed
(2) The ownership of the reserve LU is changed. (like when load balancing occurs)
(3)After the LU ownership has been changed, the LU receives a PIO related command
(other than RD/WR).
Cause There is a short period of time when changing a LU ownership is not prohibited, when a
command has not been completed yet. When the ownership of a LU is changed during that
short period of time, the buffer needed to execute the command is rewritten by the
information of another command.
Affected Products/Version DF800H/M/S, All version SA800, All version
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

11 Controller failure occurs after a host transfer DMA


Phenomena Controller failure due to DMA DCTL failure during write operations. Message displayed
when controller is blockaded: Host transfer DMA error was detected(CTL-X,DMA-XX)
Severity Medium
HDS Confidential 5 of 8
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) ISCSI interface is used.
(2)More than two write operations are issued with no read operations.
(3)DCTL error by host write transfer occurs.
Cause When a host transfer DMA DCTL failure occurs during write operations, the iSCSI chip is
not halted. So even when the DCTL has been recovered, the iSCSI chip still has been
sending data to the DCTL and a data transmission error occurs.
Affected Products/Version DF800H/M/S, All version SA800, All version
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

12 In HDP, dirty data of a virtual LU cannot be de-staged to HDD


Phenomena In HDP, dirty data of a virtual LU cannot be de-staged to HDD because the Dirty Data
Opportunity is inappropriately set. As a result, performance degradation in I/O may occur.
In the worst case, I/O operations may be suspended.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1)HDP is installed and enabled. A DP pool is set and a virtual LU is created.
(2)Either the owner cores for a virtual LU and the DPRG or the partitions to which the
CPM belong are different.
(3)The Dirty Data Opportunity exceeds the threshold for facilitating de-staging
- Performance degradation: more than 8%
- I/O suspended: more than 10% (the initial value: 5%)
(4)Any of the following behaviors are not found.
- The core for a virtual LU handles formatting the DP pool (chunks and pages) as well.
- The core for a virtual LU handles rebalancing the DP pool as well.
- The core for a virtual LU handles evacuating the HDP management area.
- There are RGs expanded in the subsystem.
Cause There are some cases where the threshold for facilitating de-staging does not apply to
virtual LU.
Affected Products/Version DF800H/M/S, V7.0/A(0870/A) or later
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

13 Controller blockade when controller executes internal retry


Phenomena Controller is blockaded because of a factor "The read data parity error was detected in the
register interface" when controller executes internal retry (CHK1 reset) in case that
controller failure occurs.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
- Intermittent failure about CPU occurs in the controller.
Cause There is a bug about illegal detection of transfer error because data transfer doesn't abort
unexpectedly when controller executes internal retry (CHK1 reset); in case that controller
failure occurs.
Affected Products/Version DF800H/M/S, All version SA8x0, All version
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

14 Controller blockade with "Micro program error [RCH]"


Phenomena When CHK1 failure occurs at destination owner's controller during the LU owner
switching, Micro program error [RCH] and controller blockade occurs.
Severity Medium

HDS Confidential 6 of 8
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) DF800H
(2) LU owner switched by the instruction opportunity or the load balancing opportunity.
Cause The process which updated a local memory in another CPU core was leaked by recovered
process of CHK1 failure. LU owner information mismatched on a cache memory and a
local memory, and it was judged as logical contradiction.
Affected Products/Version DF800H, V4.6/A(0846/A) or later
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

15 Failure of ENC firmware download in the single controller mode fails.


Phenomena After the initial setup operation in single controller mode was executed, the download of
the ENC firmware failed and the subsystem was ready (The ENC firmware is not updated).
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) Start-up in single controller mode (The single mode configuration)
(2) Execution of ENC firmware download (The ENC firmware is older than the main micro
program firmware).
Cause The Link abnormal diagnosis and the ENC firmware download operate at the same time
when LinkUp NG is detected by passing each other of internal processing before the un-
mount of the controller is detected at the start-up, and the ENC firmware download fails.
The phenomenon was generated when the Link abnormal diagnosis operates had been
changed with V8.0.
Affected Products/Version DF800H, V8.0/A(0880/A) or later
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

16 SNMP error and subsystem warning occur after firmware update.


Phenomena The SNMP error and the device (subsystem) warning state might occur after updating the
firmware. The firmware update itself completes normally. IB1700 SNMP RAM Device
allocation error for SNMP :controller /STRC
W3Q001 SNMP error occurred in the SNMP function(controller-1) :MANUAL/STRC
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) Enable SNMP Program Product
(2) Enable DHCP function
(3) Update the firmware
Cause Because the SNMP information is checked unjustified when the DHCP function is used,
and logical error has been detected.
Affected Products/Version DF800H/M/S, All version
SA8x0, All version
Fixed Product/Version Version: 0885/A or later
Category SNMP
Changed Part MICRO-Disk

17 HDU does not recover when removed and installed continuously.


Phenomena Some HDU do not recover when HDU are removed and installed continuously.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) Remove and install plural HDU continuously.
HDS Confidential 7 of 8
Cause Removing and installing plural HDU/ENC continuously, the HDU removal processing of
the firmware is executed before the previous removal processing. Then the installation
processing is executed though the backend-path is unavailable due to the removal
processing.
Affected Products/Version DF800H/M/S, V8.3/A(0883/A)
Fixed Product/Version Version: 0885/A or later
Category Common
Changed Part MICRO-Disk

18 LA error for host I/O will occur after the subsystem's recovery.
Phenomena If a Power Supply outage occurs during HDP pool recovery, LA error (Logical Address
error) for host I/O will occur after the subsystem's recovery.
Severity Medium
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) HDP Program Product is enabled
(2) Order Pool recovery during HDP pool format
(3) During pool recovery of (2), the Power Supply outage occur (without cache volatile).
(4) After subsystem recovery of (3), LA errors for host I/O will occur.
Cause There is a bug in the chunk management process of HDP. If the Power Supply is outage
(without cache volatile) during the DP pool recovery processing, some part in a chunk will
be remain as unformatted. If the host I/O accesses that chunk, LA error (logical address
error) will occur.
Affected Products/Version DF800H/M/S, V7.0/A(0870/A) or later
Fixed Product/Version Version: 0885/A or later
Category HDP
Changed Part MICRO-Disk

19 Illegal response from SSD (Solid State Disk)


Phenomena Illegal response 0B/4B06 (InitiatorResponseTimeOut) issued from SSD (Solid State Disk)
Severity Low
Conditions of Occurrence The problem may occur when all following conditions are met.
(1) SSD is included.
(2) I/O issued to SSD.
(3) Uncorrectable errors occur frequently in the SSD.
Cause The bug in the update procedure of internal information in SSD will issue illegal error
response (0B/4B06) when uncorrectable errors occur.
Affected Products/Version DF800H/M/S, V8.3/A or later
Fixed Product/Version Version: 0885/A or later
Category SSD-Firmware
Changed Part SSD-Firmware

HDS Confidential 8 of 8

You might also like