Re: State of the relay backend and rwm overlay, is it safe?

[masarati@aero.polimi.it]

> Jeffrey Crawford wrote:
>> Hello anyone ;)
>>
>> I came across this where the state of the rwm overlay was being
>> discussed. There seemed to be some indication that the state of both
>> slapd-relay and slapo-rwm "should" no longer be listed as experimental:
>>
>> http://www.openldap.org/its/index.cgi/Documentation?id=4844
>>
>> yet here we are 4 years later and both man pages still have the
>> experimental flag. In the above ticket there was some concern that the
>> configs were not yet available in cn=config and that is what was holding
>> things up.
>>
>> As far as I can tell they seem to work fine in cn=config and one of they
>> replies to the ticket mentioned that work was done in that regard, but I
>> have to wonder why the expermential flag is still present. Is there
>> still some concern? perhaps related to the following?:
>>
>> http://www.openldap.org/its/index.cgi/Incoming?id=6457
>>
>> We are trying to roll out a new LDAP service and replacing our Sun
>> Directory Server software suite. I do intend to perform testing against
>> this arch, but I'm also not 100% sure what kinds of test I should be
>> performing to ensure the rwm and relay are working correctly.
>
> I vaguely recall there are other issues with the RWM overlay but can't
> look at
> the moment. slapd-relay is fine as things go, but it's obviously useless
> without a solid RWM overlay.

As far as I recall, the main current issue is with (fancy) stacking of
overlays, which may screw up internal data restore in overlay
response/cleanup.  Specifically, there may be issues in bind and extended
operations, which are partially handled by the frontend rather than the
backend (and overlay stack).  We should streamline these calls (maybe in
OL 2.5?) such that error/success is always returned by the
backend/overlays.

p.