9:28 a.m.
I have an interest in using openldap as an external authenticator for
our membership to connect to external services. Our main table has a
primary key of a varchar (10) and nothing unique in the table that is an
int.
I would love to replace ldap_entries with a view, however, returning the
entries failed when I attempted it.
I did a bit of experimentation with the dc=example,dc=com on a search:
ldapsearch -x -b dc=example,dc=com sn=Puzdoy
The ldap_entries keyval was changed from int to varchar 10 and persons
table id was changed to varchar 10.
Below is a table on some experiments
ldap_entries.keyval and persons.id MySQL SQL SERVER 2000
2 Ok Ok
02 Ok Fail
ABC Fail Fail
abc Fail Fail
Log:
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: Constructed query: SELECT
DISTINCT ldap_entries.id,persons.id,'inetOrgPerson' AS objectClass...
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: id: '1'
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: >>> dnPrettyNormal:
<cn=Torvlobnor Puzdoy,dc=example,dc=com>
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: <<< dnPrettyNormal:
<cn=Torvlobnor Puzdoy,dc=example,dc=com>, <cn=torvlobnor puzdoy,dc=examp
...
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: backsql_oc_get_candidates():
added entry id=3, keyval=2 dn="cn=Torvlobnor Puzdoy,dc=example, ...
Jul 31 10:15:41 jpotkanski-lx slapd[30011]:
<==backsql_oc_get_candidates(): 1
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: backsql_search(): loading
data for entry id=0, oc_id=0, keyval=0
Jul 31 10:15:41 jpotkanski-lx slapd[30011]: backsql_search(): loading
data for entry id=5, oc_id=2, keyval=1
On a failed one, backsql_oc_get_candidates(): 0
openldap 2.3.39, 4.fc8 , unixodbc 2.2.12, freetds .84, mysql-odbc
latest.
Before opening anything in ITS, I wondered if this is a bug, feature
request or maybe solved in 2.4?
Jason Potkanski
Information Technology Developer
CCIM Institute
430 N. Michigan Ave, Suite 800
Chicago, IL 60611-4092
jpotkanski(a)ccim.com
www.ccim.com
tel: 312.321.8559
11:29 a.m.
----- "Jason Potkanski" <JPotkanski(a)cciminstitute.com> wrote:
I have an interest in using openldap as an external authenticator
for
our membership to connect to external services. Our main table has a
primary key of a varchar (10) and nothing unique in the table that is
an
int.
You can modify back-sql.h, defining BACKSQL_ARBITRARY_KEY and recompiling, to allow
strings to be used as keys. Not extensively tested, but should work.
p.
Ing. Pierangelo Masarati
OpenLDAP Core Team
SysNet s.r.l.
via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
-----------------------------------
Office: +39 02 23998309
Mobile: +39 333 4963172
Fax: +39 0382 476497
Email: ando(a)sys-net.it
-----------------------------------
2:05 p.m.
----- Pierangelo Masarati <ando(a)sys-net.it> wrote:
You can modify back-sql.h, defining BACKSQL_ARBITRARY_KEY and
recompiling, to allow strings to be >used as keys. Not extensively tested, but should
work.
Altered the Fedora SRPM to define BACKSQL_ARBITRARY_KEY, rebuilt the RPMs.
Works as expected with either MS SQL Server or MySQL backends. I was able to successfully
replace ldap_entries with a view and use baseObject in slapd.conf
Why isn't this #define on by default? Can this be always on or if not moved to a
configuration variable in slapd.conf to turn it on?
Jason Potkanski
Information Technology Developer
CCIM Institute
430 N. Michigan Ave, Suite 800
Chicago, IL 60611-4092
jpotkanski(a)ccim.com
www.ccim.com
tel: 312.321.8559
2:13 p.m.
----- "Jason Potkanski" <JPotkanski(a)cciminstitute.com> wrote:
----- Pierangelo Masarati <ando(a)sys-net.it> wrote:
>You can modify back-sql.h, defining BACKSQL_ARBITRARY_KEY and
recompiling, to allow strings to be >used as keys. Not extensively
tested, but should work.
Altered the Fedora SRPM to define BACKSQL_ARBITRARY_KEY, rebuilt the
RPMs.
Works as expected with either MS SQL Server or MySQL backends. I was
able to successfully replace ldap_entries with a view and use
baseObject in slapd.conf
Why isn't this #define on by default? Can this be always on or if not
moved to a configuration variable in slapd.conf to turn it on?
Because it was a hack, initially not meant for release; because allowing the two
functionalities run-time is sort of a mess; because using integers is way more efficient.
Probably, it's something to consider anyway (or, at least, to document).
p.
Ing. Pierangelo Masarati
OpenLDAP Core Team
SysNet s.r.l.
via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
-----------------------------------
Office: +39 02 23998309
Mobile: +39 333 4963172
Fax: +39 0382 476497
Email: ando(a)sys-net.it
-----------------------------------
2:50 p.m.
--On Friday, August 01, 2008 4:05 PM -0500 "Potkanski, Jason"
<JPotkanski(a)cciminstitute.com> wrote:
----- Pierangelo Masarati <ando(a)sys-net.it> wrote:
> You can modify back-sql.h, defining BACKSQL_ARBITRARY_KEY and
> recompiling, to allow strings to be >used as keys. Not extensively
> tested, but should work.
Altered the Fedora SRPM to define BACKSQL_ARBITRARY_KEY, rebuilt the
RPMs.
Works as expected with either MS SQL Server or MySQL backends. I was able
to successfully replace ldap_entries with a view and use baseObject in
slapd.conf
Why isn't this #define on by default? Can this be always on or if not
moved to a configuration variable in slapd.conf to turn it on?
Can't you just set it in CFLAGS when you compile?
--Quanah
--
Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc
--------------------
Zimbra :: the leader in open source messaging and collaboration
11:27 p.m.
Can't you just set it in CFLAGS when you compile?
Actually, in CPPFLAGS. But I don't think so, as there's an explicit #undef FLAG
in back-sql.h (unless -DFLAG overrides it, dunno). That's what I mean by
"exposing" it: allow it to be set without hacking the code.
p.
Ing. Pierangelo Masarati
OpenLDAP Core Team
SysNet s.r.l.
via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
-----------------------------------
Office: +39 02 23998309
Mobile: +39 333 4963172
Fax: +39 0382 476497
Email: ando(a)sys-net.it
-----------------------------------
4875
days inactive
4877
days old
openldap-software@openldap.org
5 comments
3 participants
participants (3)
-
Pierangelo Masarati -
Potkanski, Jason -
Quanah Gibson-Mount