Jeff,
KPID consistency checks out (i won't bother to post details here). But, the same KPID is involved in all of the wait data. And no, i'm not better than that. never thought to check it.
The disclosure here is that I am using the "asemon" tool to collect MDA data into it's repository. So, I do have raw data, but it's the data as collected by the "asemon_logger" component. I am certainly suspecting that the delay is in reading the TEXT chain.
Question: Can such a long delay be due to forcing other dirty pages past wash marker in cache and waiting for IO to complete on those? if so, wouldn't I see those kinds of wait events too?
I have a feeling that my mda collection is a bit corrupt here, and i'm not getting the whole picture.