Perhaps the cause is the obvious:  RACDCERT has some options which require
ICSF.  For example a certificate that is migrated from one RACF system to
another will have PKDS.  So fairly easy to wind up using ICSF unexpectedly.

On Sat, Jun 22, 2024 at 8:12 AM Phil Smith III <[email protected]> wrote:

> (Cross-posted to IBMTCP-L and IBM-MAIN)
>
> Had an odd one this morning: a customer who was doing some testing could
> not connect to our server (on premises at their site) from z/OS (server is
> an x86 Linux machine). I saw the email when I woke up and thought "OK,
> gsktrace to the rescue!"
>
> But by the time I got to my desk, I had more email saying "Nevermind, ICSF
> wasn't running--once we started it, all is fine". And now that's working,
> they can't break it again to run with gsktrace.
>
> Meanwhile, I can connect just fine without ICSF running. Of course, that's
> to one of OUR versions of the same server, using one of OUR certificates.
> Wild guess is that the customer's cert is using some certificate feature
> that requires ICSF interpretation, but I had them send me both the root and
> the leaf, and various online cert analyzers don't show anything obvious.
>
> Anyone know of any certificate features that absolutely require ICSF
> intervention? Our product uses GSK directly -- no AT-TLS or anything like
> that.
>
> I realize this is vague but hoping someone (maybe at IBM?) has a guess...
>
> Thanks,
> ...phsiii
>
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN
>

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to