IBM Support

IV25130: MISSING NBTNS QUERY ACTION TO RESOLVE FOR THE IP ADDRESS OF THE DOMAIN CONTROLLER CAUSING A SCAN TO SKIP WIN-MS-KB CHECKS.

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as fixed if next.

Error description

  • [Symptom]
    Extracted from Mercury.
    
    ES750 - difference results between 2 scans on same target
    
    [Detailed Description]
    Two scans were run using a ES750 on a Windows 2008 target but
    the results are different.  One scan runs Windows missing
    patches checks while the other does not.  From the provinfo, it
    appears in the case where Windows missing patches checks are
    run, the ES is able to connect to the Domain Controller directly
    to verify the assessment credentials.  While in the other case,
    it could not contact the DC, and subsequently the Windows
    missing patches checks are skipped.  But based on L3
    documentation, the verification of account should strictly be
    performed on the target and not on the DC.  This CR is opened to
    investigate and confirm the correct expected behavior.
    
    [Additional Information]This is the PMR that we discussed where
    the ES actually connected to the DC directly to verify
    credentials.  The snippet are in the file below:
    
    [attachment "analysis1.txt" deleted by Cheston S
    Chan/Atlanta/IBM]
    
    Note the lines preceeded by >>>>>.
    
    The provinfo can also be downloaded from the link below:
    
    ftp://pokgsa/home/c/s/cschan/web/public/44766.000.738.provinfo_I
    TD4SCAN1Z.whbhk.net_123.87.11.105_20120320.tgz
    
    Please let me know what you can derive after your review.
    
    ==========
    Lynn Lowrie <llowrie>, 6/18/2012:
    
    Reviewing the data provided.
    ~
    ---End Comment Update 6/18/2012 5:32:04 PM---
    ________________________________________
    Cheston Chan <cchan>, 6/27/2012:
    Do you have any updates on this CR?
    ---End Comment Update 6/27/2012 9:37:36 AM---
    Attaching my email to L3 here for reference:
    
    Do you have any updates on this CR?   As a reminder, the
    customer is seeing the behavior that sometimes the Windows patch
    checks are skipped if verifyAdminCredentials is enabled in the
    policy when scanning Windows 2008 targets.
    
    Upon further review of the logs and pcap, I discovered some
    interesting facts that I would like to share.  It appears in the
    case of failing the verifyAdminAccount, the ES scanner does not
    run an NBNS name query to resolve for the domain ITD_DOMAIN.  I
    have attached my analysis here to see if you can find the logic
    in the code as to why the ES skipped the NBNS call.
    
    Scan sequence A then B then C then D:
    
    A> Windows Server 2003 at 10.100.0.109
    
    1. During the discovery scan of the Windows Server 2003 target,
    the ES resolved the domain ITD_DOMAIN to IP address through NBNS
    name query
    
    
    2. In the assessment scan, the ES used the cached DC IP address
    to run the scan.  And it contacted the DC to verify the user
    account credentials.
    ES ran the Windows patch checks because verifyAdminAccount
    yielded true.
    
    B> Windows Server 2008 at 10.100.33.6
    
    Both discovery and assessment scans of the Windows Server 2008
    target used the cached domain IP address to verify credentials
    acquired from A>.
    Associated pcap does not show any packet activities of another
    NBNS call so it is most likely using the cached value.
    
    
    ES ran Windows patch checks because verifyAdminAccount yielded
    true using cached domain IP address acquired in A>
    
    C> **** es stop/start scripts were run on the ES scanner.   ****
    
    D> Windows Server 2008 at 10.100.33.5
    
    1. The es stop/start or the expiration of time (10 minutes?)
    removed the cached domain IP address.
    
    2. During the discovery scan, the ES did not send NBNS name
    query to resolve the domain IP address.
    
    
    3. During the assessment scan, the discovery could not resolve
    for the DC IP address but it did not issue NBNS call so domain
    was not resolved.
    
    
    As a result, it skipped the Windows missing patches checks
    because verifyAdminAccount yielded false.
    
    Therefore, the issue is caused by the ES skipped running NBNS
    query to resolve the IP address of the domain NB ITD_DOMAIN.
    
    Please let me know if you have any suggestions.
    ---End Comment Update 7/3/2012 2:07:35 PM---
    ________________________________________
    Lynn Lowrie <llowrie>, 7/13/2012:
    
    Found what we were looking for as far as when we access the CD
    during Account Verification.
    
    If the "Verify account access level before using" check box is
    checked, we check on the target Local and see if the account
    returns as admin. If it does, we see no need to check the domain
    controller for a local account. If the, "Check local group
    membership to verify access level" check box is also enabled, we
    do a check of the groups and users in the Administrators group.
    If the account in question is not found and the, "Access domain
    controllers to verify access level" check box is enabled, then
    we check the domain controller.
    
    First we do an NbtNs call to locate the DC and then we will
    query. We also have functionality that will remember when a
    successful admin account has been found, but I cannot find the
    timeout of when that becomes lost. But when admin level has been
    determined and the account marked as verified, we will remember
    the smb resource for subsequent calls to verifyAccount.
    
    In this case when the call is made to the DC to verify, it has
    passed from the smb resource buffer and the query had to be made
    to verify the admin level.
    ~
    ---End Comment Update 7/13/2012 12:53:58 AM---
    ________________________________________
    
    Cheston Chan <cchan>, 7/20/2012:
    Hi Lynn,
    
    It is great to hear from you.
    
    Thank you for the detail explanation about the logic on
    verifyAdminAccount. Your explanation does match to the scenario
    as extracted from the provinfo, except for points C and D,
    referencing back to the "Scan sequence A then B then C then D"
    in my previous post.  The issue occurred after an es stop/start.
    The fact that the ES could not resolve the DC IP address after
    the es restart tells me that the cached value is most likely
    erased and a new NBNS query should have been performed.
    
    Why did the ES not even attempt to issue a NBNS command to find
    the DC to derive its IP address after it failed to locate the
    user account in the Administrators group on the target, and then
    went ahead to conclude that the Domain cannot be resolved?
    
    Based on your explanation that if "Access domain
    controllers to verify access level" is TRUE, then the ES should
    contact the DC by NBtNS.  Yet it appeared to have skipped this
    action.  Can you look further in the code and see if there are
    any criteria to run the NbtNS logic, which might have caused the
    action to be skipped?
    
    I had gone ahead and pointed out to customer that the account is
    missing in the local Administrator and provided the needed steps
    to add in the account, and waiting for customer's results.  But
    it would be great if the ES can consistently contact the DC if
    the flag is enabled.
    
    ---End Comment Update 7/20/2012 4:38:25 PM---
    ---------------------------------------
    Cheston Chan <cchan>, 7/24/2012:
    Customer's policy does not permit adding another local admin
    account to a target that has already been added to the domain.
    We need to find out why the ES scanner skipped the NbtNS query
    when the logging showed that it is resolving for the IP address
    of the domain controller.
    Do you have any updates from your research in the code?
    ---End Comment Update 7/24/2012 8:13:32 AM---
    

Local fix

  • Customer rejected workaround by adding a local admin account to
    the target Administrators group due to his security policy.
    

Problem summary

  • Newer Microsoft Operating Systems utilize restricted SAMR
    communication models not implemented in the current release of
    Enterprise Scanner 2.3.
    
    This cannot be addressed in a patch and will have to be
    addressed when a new project is slated.
    

Problem conclusion

Temporary fix

Comments

APAR Information

  • APAR number

    IV25130

  • Reported component name

    PROVENTIA ES750

  • Reported component ID

    5121211FW

  • Reported release

    230

  • Status

    CLOSED FIN

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2012-07-24

  • Closed date

    2012-09-27

  • Last modified date

    2012-09-27

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

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

Fix information

Applicable component levels

  • R230 PSY

       UP

[{"Business Unit":{"code":"BU008","label":"Security"},"Product":{"code":"SSETCY","label":"Proventia Network Enterprise Scanner"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"230","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
27 September 2012