On (17/01/25 09:03), Sumit Bose via sssd-users wrote:
>Am Thu, Jan 16, 2025 at 01:17:56PM -0600 schrieb Spike White via sssd-users:
>> Alexey,
>> 
>> Thank you.  That seems right.  But the openldap client lib content doesn't
>> seem like it's being written to the /var/log/sssd/*.log files.
>> 
>> I'm able to turn on debug_level = 5  in the domain section of the sssd.conf
>> file and see the fully-qualified user names requested.  But  even setting
>> 
>> ldap_library_debug_level = -1
>> 
>> seems to yield no additional LDAP content in /etc/sssd/*.log files.
>> 
>> Is it because my id_provider and auth_provider is ad and not ldap?  Or does
>> the openldap client lib write its output to another location?
>
>Hi,
>
>no, unfortunately the man page (or the code) is wrong,
>https://github.com/SSSD/sssd/blob/master/src/providers/ldap/sdap.c#L840
>currently you need `debug_level = 9` to get the libldap debug logs.
>

`debug_level = 0x4000` should provide less logs than `debug_level = 9` :-)
https://github.com/SSSD/sssd/blob/85784e7614da54aeaed2e8a757b5bd3c9523125d/src/util/debug.h

LS

>I think it would be useful to have the behavior as described in the man
>page, are you interested to provide a patch?
>
>bye,
>Sumit
>
>> 
>> Spike
>> 
>> On Thu, Jan 16, 2025 at 11:45 AM Alexey Tikhonov <[email protected]>
>> wrote:
>> 
>> > Hi,
>> >
>> > I think this is `man ldap_set_option :: LDAP_OPT_DEBUG_LEVEL`:
>> > ```
>> > Valid debug levels are LDAP_DEBUG_ANY, LDAP_DEBUG_ARGS, LDAP_DEBUG_BER,
>> > LDAP_DEBUG_CONNS, LDAP_DEBUG_NONE,  LDAP_DEBUG_PACKETS,  LDAP_DEBUG_PARSE,
>> >  and  LDAP_DEBUG_TRACE
>> > ```
>> >
>> > For values see
>> > https://git.openldap.org/openldap/openldap/-/blob/master/include/ldap_log.h?ref_type=heads#L107
>> >
>> >
>> > On Thu, Jan 16, 2025 at 5:09 PM Spike White via sssd-users <
>> > [email protected]> wrote:
>> >
>> >> Sssd experts,
>> >>
>> >>
>> >>
>> >> Our AD team is complaining that their AD controllers in our more
>> >> densely-populated datacenters are getting overloaded.  (up to 5K 
>> >> concurrent
>> >> LDAP connections per AD DC).
>> >>
>> >>
>> >> They find that 90% of the LDAP queries come from Linux servers -- which
>> >> is surprising as Linux servers are < 50% of the resident population.
>> >>
>> >>
>> >> We had previously found a coding bug in a vendor’s cron job that was
>> >> generating a lot of unnecessary LDAP queries.  That cybersecurity
>> >> monitoring software was on every Linux server, so it led to a thundering
>> >> herd problem when the cron job would kick off on every Linux server.
>> >>
>> >>
>> >> That thundering herd problem has been remediated, but we suspect
>> >> something similar occurring now.  We’d like to see what LDAP queries are
>> >> being executed by several random Linux servers, but we’ve had negative
>> >> experiences leaving sssd debug_level = 9 on for several days.  (Fills up
>> >> the /var/log filesystem after a few days).
>> >>
>> >>
>> >> I read in the sssd-ldap man page:
>> >>
>> >>
>> >>        ldap_library_debug_level (integer)
>> >>
>> >> Switches on libldap debugging with the given level. The libldap debug
>> >> messages will be written independent of the general debug_level.
>> >>
>> >>
>> >>
>> >> OpenLDAP uses a bitmap to enable debugging for specific components, -1
>> >> will enable full debug output.
>> >>
>> >>
>> >>
>> >>            Default: 0 (libldap debugging disabled)
>> >>
>> >>
>> >>
>> >> This seems to be exactly what we want!  We want to see what LDAP queries
>> >> are being executed without all the other massive logging that arises from
>> >> debug_level 9.
>> >>
>> >>
>> >>
>> >> However, I cannot find these bitmap values of debug levels for the
>> >> openldap library.  The most I can find is bitmapped debug values for a
>> >> stand-alone slapd daemon.
>> >>
>> >>
>> >>
>> >> OpenLDAP debug level - Knowledge Base / Community - Univention Help
>> >> <https://help.univention.com/t/openldap-debug-level/19301>
>> >>
>> >>
>> >>
>> >> Where may I find the documentation of the bitmapped debug values that can
>> >> be set via ldap_library_debug_level?
>> >>
>> >>
>> >>
>> >> Spike White
>> >>
>> >>
>> >> --
>> >> _______________________________________________
>> >> sssd-users mailing list -- [email protected]
>> >> To unsubscribe send an email to [email protected]
>> >> Fedora Code of Conduct:
>> >> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>> >> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>> >> List Archives:
>> >> https://lists.fedorahosted.org/archives/list/[email protected]
>> >> Do not reply to spam, report it:
>> >> https://pagure.io/fedora-infrastructure/new_issue
>> >>
>> >
>
>> -- 
>> _______________________________________________
>> sssd-users mailing list -- [email protected]
>> To unsubscribe send an email to [email protected]
>> Fedora Code of Conduct: 
>> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>> List Archives: 
>> https://lists.fedorahosted.org/archives/list/[email protected]
>> Do not reply to spam, report it: 
>> https://pagure.io/fedora-infrastructure/new_issue
>
>-- 
>_______________________________________________
>sssd-users mailing list -- [email protected]
>To unsubscribe send an email to [email protected]
>Fedora Code of Conduct: 
>https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>List Archives: 
>https://lists.fedorahosted.org/archives/list/[email protected]
>Do not reply to spam, report it: 
>https://pagure.io/fedora-infrastructure/new_issue

-- 
_______________________________________________
sssd-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedorahosted.org/archives/list/[email protected]
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to