On Sun, Dec 24, 2017 at 02:21:58PM +0000, Francis Dupont via RT wrote:
> BTW on Fedora 27 with OPENSSL_FORCE_FIPS_MODE set to 1
> in the environment (easier than to boot in FIPS mode):
> 
> ./named -g -c /dev/null
> 24-Dec-2017 15:17:29.432 starting BIND 9.12.0rc1 <id:unset_id>
> 24-Dec-2017 15:17:29.432 running on Linux x86_64 4.14.7-300.fc27.x86_64 #1 SMP Mon Dec 18 16:06:12 UTC 2017
> 24-Dec-2017 15:17:29.432 built with '--enable-full-report' '--enable-developer'
> 24-Dec-2017 15:17:29.432 running as: named -g -c /dev/null
> 24-Dec-2017 15:17:29.432 ----------------------------------------------------
> 24-Dec-2017 15:17:29.432 BIND 9 is maintained by Internet Systems Consortium,
> 24-Dec-2017 15:17:29.432 Inc. (ISC), a non-profit 501(c)(3) public-benefit 
> 24-Dec-2017 15:17:29.432 corporation.  Support and training for BIND 9 are 
> 24-Dec-2017 15:17:29.432 available at https://www.isc.org/support
> 24-Dec-2017 15:17:29.432 ----------------------------------------------------
> 24-Dec-2017 15:17:29.432 found 2 CPUs, using 2 worker threads
> 24-Dec-2017 15:17:29.432 using 1 UDP listener per interface
> 24-Dec-2017 15:17:29.432 using up to 4096 sockets
> 24-Dec-2017 15:17:29.433 md5.c:58: fatal error:
> 24-Dec-2017 15:17:29.433 RUNTIME_CHECK(EVP_DigestInit(ctx->ctx, EVP_md5()) == 1) failed
> 24-Dec-2017 15:17:29.433 exiting (due to fatal error in library)
> Abort (core dumped)

I recommend printing a user-friendly message than the above when
EVP_md5() cannot be used, i.e., print an explanation message than using
RUNTIME_CHECK().

		Mukund