IBM Support

IT34081: DDT TABLE GROWING AGAIN AFTER MASS EVICTION EVEN WITH FIX FOR IT32392

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • With APAR IT32392, the mass eviction process was updated to
    ensure the parameter used to control the vSnap Deduplication
    Table (DDT) size was kept active after reboot even if the DDT
    monitored internal threshold value was slightly above the hard
    coded value.
    It has been seen that the tolerance currently
    coded is too small and that, on some occasions, the DDT size
    control parameter again was disabled unexpectedly after vSnap
    reboot.
    
    IBM Spectrum Protect Plus Versions Affected:
    IBM
    Spectrum Protect Plus 10.1.5 e-fix4 and above
    IBM Spectrum
    Protect Plus 10.1.6
    
    | MDVPARTL 10.1.5 e-fix4 TIV_5737SPLUS |
    IT32392
    | MDVPARTL 10.1.6 TIV_5737SPLUS | IT32392
    
    Initial
    Impact: Medium
    
    Additional Keywords: SPP, SPPLUS, TS003935044,
    mass eviction, DDT, dedup
    

Local fix

  • Until the fix is available, contact IBM Support to proceed with
    the mass eviction process.
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * IBM Spectrum Protect Plus level 10.1.5 and 10.1.6.           *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * See Error Description                                        *
    ****************************************************************
    * RECOMMENDATION:                                              *
    * Apply fixing level when available. This problem is currently *
    * projected to be fixed in IBM Spectrum Protect Plus level     *
    * 10.1.7. Note that this is subject to change at the           *
    * discretion of IBM.                                           *
    ****************************************************************
    

Problem conclusion

  • Starting with the fixes introduced in IT31918, vSnap uses a
    kernel module parameter to control the amount of unique entries
    in the DDT. The parameter is set dynamically on startup and it
    sets a threshold on the amount of unique entries. In the normal
    course of operations as new writes occur, the number of entries
    can grow slightly over this threshold. A background task is
    responsible for monitoring the growth and pruning entries to
    bring the number back to the threshold value.
    
    Under abnormal conditions (e.g. when the vSnap server is
    experiencing I/O hangs due to issues described in IT31841 or
    IT34064), the background task is not able to prune unique
    entries, so the number grows beyond the threshold. The next time
    the vSnap is rebooted, it checks the current number of unique
    entries. If the current number is greater than [threshold +
    additional headroom], vSnap deliberately disables the threshold.
    This is by design, because if the headroom is too large, then a
    large amount of entries would need to be pruned immediately
    which would result in further slowdowns. Instead, the solution
    in this case is to manually perform a mass eviction of unique
    entries at a later convenient time.
    
    If the headroom is too small, the threshold gets disabled even
    when the number of unique entries is only slightly above the
    threshold.
    
    To resolve this issue, the default value of the headroom has
    been increased. The value of the headroom is also configurable
    via a vSnap preference named ddtUniqueMaxHeadroom.
    

Temporary fix

Comments

APAR Information

  • APAR number

    IT34081

  • Reported component name

    SP PLUS

  • Reported component ID

    5737SPLUS

  • Reported release

    A15

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2020-09-02

  • Closed date

    2020-09-25

  • Last modified date

    2020-09-25

  • APAR is sysrouted FROM one or more of the following:

  • APAR is sysrouted TO one or more of the following:

Fix information

  • Fixed component name

    SP PLUS

  • Fixed component ID

    5737SPLUS

Applicable component levels

[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSNQFQ","label":"IBM Spectrum Protect Plus"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"A15","Line of Business":{"code":"LOB26","label":"Storage"}}]

Document Information

Modified date:
31 January 2024