Open 389-ds-bot opened 4 years ago
Comment from firstyear (@Firstyear) at 2018-04-05 14:41:34
I'm on the fence about this. On one hand the standard says all backends should be in the namingContexts section for discoverability.
On the other, these are internal details of the server, which are not real backends - they are "implementation details". Think of it as the "operational attribute" of backends.
I'm erring to standards correctness however, because I think we should do the "right thing", as obscurity does not add security or other niceness, and this is generally only use by applications like apache dirstudio.
Comment from firstyear (@Firstyear) at 2018-04-05 14:41:35
Metadata Update from @Firstyear:
Comment from tbordaz (@tbordaz) at 2018-04-05 15:43:49
In the sense of https://tools.ietf.org/html/rfc4512#section-5, cn=config and cn=monitor being root of subordinates entries are namingcontext of the server. So I tend to agree with the proposal. I do not see special difficulty to implement it. The only non namingcontext being the root DSE.
Comment from mreynolds (@mreynolds389) at 2018-04-05 17:38:18
Metadata Update from @mreynolds389:
Hi, is there an update on this?
Cloned from Pagure issue: https://pagure.io/389-ds-base/issue/49630
Tools like Apache Directory Studio use namingContexts on root DSE to show all available trees. In 389-DS, the multi-valued attribute doesn't contain
cn=config
andcn=monitor
.Please consider to include both in-memory databases in namingContexts. The lack of
cn=config
in namingContexts makes rather inconvenient to explore 389-DS configuration with Apache Directory Studio.