Understanding Data Connectors

[Kanuch, Andrew]

Today I resolved a (self-inflicted) slowness problem on my IDP, but I am confused as to why the problem was happening in the first place. This is only for personal education--I don't have any technical issues at this point.

Previously I had tried using a recursive group member search ( from https://wiki.shibboleth.net/confluence/display/SHIB2/ResolverScriptAttributeDefinitionExamples) for a project and it did not work. (With the difference that I increased the SearchTimeLimit up to 15 seconds.)

But I left the second data connector in my attribute resolver XML file . One attribute definition (The one in the example) had the data connector as a dependency.  The attribute definition was not in any attribute-filters (which I realize in hindsight probably doesn't matter, and I should treat the Filter as just that.)

I removed the second data connector and it's attribute definition from the attribute-resolver XML and bam, my IDP login experience is down to a few seconds.

What I am confused on is this: Was this triggered by the attribute definition, or the data connector? If I had left the data connector in, and simply removed the Attribute Definition that listed it as a dependency, would it still trigger the 15 seconds?

Side Note: I saw the Shib3 briefing on Wednesday from InCommon, and I'm pumped! I hope there is hosted training soon.