Scribd Logo
Upload

Welcome to Scribd!

  • 8 views

    9200 X5xme

    Uploaded by

    asdf
    This document summarizes improvements and changes made in microcode version 055M/E, 355M/E, and 050M/E for an HDS storage subsystem. Six improvements are described: 1) enhancing a data transfer check function, 2) improving vulnerability to LAN attacks, 3) adding a setting for auto diagnosis thresholds, 4) improving phase I of backend auto diagnosis, 5) improving phase II of backend auto diagnosis, and 6) improving handling of memory obstacles on enclosure controllers. Six changes are also noted: 1) fixing a data share command timeout issue, 2) ensuring write data is flushed to HDDs, 3) addressing a subsystem shutdown from a single HDD failure, 4) fixing host connection failures

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd

    9200 X5xme

    Uploaded by

    asdf
    8 views3 pages
    This document summarizes improvements and changes made in microcode version 055M/E, 355M/E, and 050M/E for an HDS storage subsystem. Six improvements are described: 1) enhancing a data transfer check function, 2) improving vulnerability to LAN attacks, 3) adding a setting for auto diagnosis thresholds, 4) improving phase I of backend auto diagnosis, 5) improving phase II of backend auto diagnosis, and 6) improving handling of memory obstacles on enclosure controllers. Six changes are also noted: 1) fixing a data share command timeout issue, 2) ensuring write data is flushed to HDDs, 3) addressing a subsystem shutdown from a single HDD failure, 4) fixing host connection failures

    Original Description:

    test

    Original Title

    9200 X5XME

    Copyright

    © © All Rights Reserved

    Available Formats

    DOC, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes improvements and changes made in microcode version 055M/E, 355M/E, and 050M/E for an HDS storage subsystem. Six improvements are described: 1) enhancing a data transfer check function, 2) improving vulnerability to LAN attacks, 3) adding a setting for auto diagnosis thresholds, 4) improving phase I of backend auto diagnosis, 5) improving phase II of backend auto diagnosis, and 6) improving handling of memory obstacles on enclosure controllers. Six changes are also noted: 1) fixing a data share command timeout issue, 2) ensuring write data is flushed to HDDs, 3) addressing a subsystem shutdown from a single HDD failure, 4) fixing host connection failures

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    8 views3 pages

    9200 X5xme

    Uploaded by

    asdf
    This document summarizes improvements and changes made in microcode version 055M/E, 355M/E, and 050M/E for an HDS storage subsystem. Six improvements are described: 1) enhancing a data transfer check function, 2) improving vulnerability to LAN attacks, 3) adding a setting for auto diagnosis thresholds, 4) improving phase I of backend auto diagnosis, 5) improving phase II of backend auto diagnosis, and 6) improving handling of memory obstacles on enclosure controllers. Six changes are also noted: 1) fixing a data share command timeout issue, 2) ensuring write data is flushed to HDDs, 3) addressing a subsystem shutdown from a single HDD failure, 4) fixing host connection failures

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    of 3

    9200

    MICROCODE VERSION 055M/E, 355M/E & 050M/E


    RELEASED 10/03/2006

    Newly supported features/functions Version 055M/E, 355M/E & 050M/E

    NONE

    Improvements for Version 055M/E, 355M/E & 050M/E

    1) Check function improvement for “Data transfer check error”


    Item Description
    Contents The LBA check function of data transfer from the cache memory to the HDD is improved.
    When an LBA error is detected by the additional check, the controller displays the following
    message: WEB message: “HF010E Data transfer check error [DSTPOSERR]”

    2) Improvement for vulnerability to LAN attack


    Item Description
    Contents The LAN communication from Web browser to the Subsystem may be blocked by LAN attack.
    The microcode is improved to resolve this problem.

    3) Backend Auto Diagnosis Threshold setting support


    Item Description
    Contents The microcode supports a function to make the maximum retry count of the backend auto
    diagnosis a changeable setting. The default entry is set at 10.

    4) Improvement of Backend Auto Diagnosis (Phase I)


    Item Description
    Contents The microcode is improved to prevent an incorrect judgment of a failed part by an intermittent
    obstacle in the backend loop.

    5) Improvement of Backend Auto Diagnosis (Phase II)


    Item Description
    Contents The microcode is modified to improve the ability to isolate an intermittent obstacle component
    of the backend loop by increasing the number of diagnostic frames up to a max. Of 15 times to
    check the backend loop.

    6) Improvement of the subsystem down by memory obstacle of an ENC


    Item Description
    Contents When a memory failure of ENC occurs, subsystem down may occur due to a flaw in the logic
    that determines that both ENC have a problem.
    The microcode is modified to improve the check process of ENC memory in this case.

    HDS Confidential 1 of 3
    The change of contents for Version 055M/E, 355M/E & 050M/E

    1) Data share command cannot be cleared by a timeout


    Item Description
    Contents The timeout procedure may not work correctly at the data share command time-out, and the data
    share cannot be cleared by the time-out. The other host commands or internal processes that
    should be executed have to wait. After the cache hit data transfer to the host, the microcode sets
    an internal flag and this flag remains until the other data transfer to/from the host occurs. The
    timeout procedure in the microcode incorrectly refers to this flag and cancels the time-out
    procedure. The problem occurs when all of the following conditions are met:
    A) Data share job takes a long time to complete
    B) The data share job time-out is detected just after the cache hit data transfer to the host and no
    other data transfer occurs

    2) Write data in the cache memory is not written to the HDD.


    Item Description
    Contents The command access from the host to the subsystem may fail (time-out). To write data to HDD,
    the microcode periodically examines whether dirty data exists in the cache memory or not.
    However the microcode may not check the dirty data correctly for the HDD located after
    number 65. The problem occurs when all of the following conditions are met:
    A) LU are defined on 65 or more HDD
    B) There is no dirty data from 0~64th HDD
    C) Dirty data in HDD after 65th, (more segments than threshold value) in cache are used
    D) There are no internal jobs to release a segment in cache memory

    3) Subsystem down by one SES HDD failure


    Item Description
    Contents The problem occurs when all of the following conditions are met:
    A) Dual controller configuration
    B) Communication of one SES HDD 0,1,2,3 and ENC fails
    C) The HDD has permanent obstacles on both loops
    Or
    A) Dual controller
    B) ENC problem with an abnormal SES function occurs
    C) In case of ECN#0, the even HDD has a permanent obstacle on both loops
    Or In case of ENC#1, the odd HDD has a permanent obstacle on both loops

    4) Host connection failure due to HDD obstacle


    Item Description
    Contents A host may fail to access the LU after an HDD is blocked
    The problem occurs when all of the following conditions are met:
    A) HDD command time-out is detected by online verify processing
    B) A command from the LAN such as DAMP/SNM configuration setting or login process for
    password protection start execution just after the above time-out

    HDS Confidential 2 of 3
    5) LU Hang up
    Item Description
    Contents LU may hang because the controller fails to block when an obstacle occurs. This is caused by a
    bug in the microcode logic to block the controller when a specific obstacle occurs on the
    controller. The problem occurs when all of the following conditions are met:
    A) The specific obstacle occurs in the controller
    B) LAN access occurs frequently

    6) Subsystem down with “Backend down [NO SES-HDU BKD]


    Item Description
    Contents Subsystem down may occur after FC HDD block. In HDD obstacle detection, microcode
    requires Bypass operation of obstacle HDD over the ENC by SES access. Because microcode
    treated an ENC obstacle as a SES HDD obstacle incorrectly when Bypass instructions failed,
    subsystem down can occur. Problem occurs with the following conditions met:
    A) FC HDD block
    B) Bypass processing for the obstacle HDD fails due to an ENC obstacle
    C) Retry of bypass processing fails

    HDS Confidential 3 of 3

    You might also like