[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Schema design questions



> Hello,
>
> We are in the process of converting a large LDAP directory running
> OpenLDAP
> 1.2 to OpenLDAP 2.3,

good luck ;)

> and since our current schema is quite custom, we need
> to recreate it for the new syntax. So far a first try is looking good, but
> I
> have a few questions for which I have not found much documentation :
>
> - How useful/needed are the default schemas (core.schema, openldap.schema,
> ...) for slapd operations ? Since our schema is not at all based on x500
> attributes/objects, I have not included them, and it seems to work fine,
> but
> are there any potential issues to be aware of ? (I noticed slapd defines a
> few base objects/attributes internaly, is that enough ?)

Those that are __strictly__ required for operating are actually hardwired
in slapd; few more are recommended (e.g. groupOfNames/member are not
hardwired, so you need to include the appropriate files if you plan to use
group ACLs or so).  They don't do any harm, as far as I know, unless you
consider harm having few dozens extra entries in a couple rbtrees that are
searched quite often, and the related performance impact.

>
> - Is there any difference between UTF8 and IA5 string attribute types
> other
> than the obvious syntax checks on writes, such as performance difference
> on
> searches perhaps ?

IA5 => ~ 0 -> 127 (ASCII)
UTF8: ~ any

IA5 and UTF8, with respect to searches, only differ in the
validation/normalization; after that, comparisons occur by way of exactly
the same functions.  When the subset of UTF8 that corresponds to IA5 is
used (the vast majority of the cases, for English speaking users, and a
geat deal of cases for everyone else; apparently your name is an exception
to this ;), there is no difference at all.

>
> - Not strictly related to the schema itself, but are there any issues
> running bdb/hdb with an empty suffix ?

None that I'm aware of.  The empty suffix is consistently checked in
test027 since 2.2.something and, apart from few details when adding or
renaming a direct subordinate, or when using "" as the searchBase, it does
not use any special code, so in principle there shouldn't be any issue. 
If there are, they're a bug, so please report them thru the ITS.

p.

-- 
Pierangelo Masarati
mailto:pierangelo.masarati@sys-net.it



Ing. Pierangelo Masarati
Responsabile Open Solution

SysNet s.n.c.
Via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
------------------------------------------
Office:   +39.02.23998309          
Mobile:   +39.333.4963172
Email:    pierangelo.masarati@sys-net.it
------------------------------------------