Upon further investigation, this seems to be caused by the --oval-results
option. And yes, I realized that a) I don't need the
--check-engine-results options and b) I can create a tailoring file for use
without the datastream. So regardless of whether or not I use the
datastream or the seperate xccdf and cpe files, adding the --oval-results
option kills the report generation. (the machine has 16GB of RAM btw).
Also, should I move the discussion to the Red Hat GovSec list instead?
On Tue, Oct 6, 2015 at 12:40 PM, Chuck Atkins <chuck.atkins(a)kitware.com>
I just did a fresh install of RHEL6, with minimal desktop
applied all updates, and installed scap-security-guide, openscap, and
scap-workbench from the system repos. First I created a tailoring file
with SCAP Workbench to adjust the STIG server profile for desktop usage
(allow X, remove various server packages, etc.). The scan worked fine but
when generating the results and report, I get a bunch of memory allocation
errors. When I tried to evaluate from the command line instead, using the
tailoring file from scap-workbench, I get:
[root@rhel6-stig-vm ~]# oscap xccdf eval \
--tailoring-file ssg-rhel6-kw-desktop-xccdf.xml \
--profile xccdf_kitware.com_profile_stig-rhel6-server-upstream_desktop \
--report ssg-report.html --results ssg-results.xml \
--check-engine-results --oval-results \
all benchmark rules process
XPath error : Memory allocation failed : growing nodeset hit limit
growing nodeset hit limit
runtime error: file /usr/share/openscap/xsl/xccdf-report-oval-details.xsl
line 39 element key
Failed to evaluate the 'match' expression.
(error repeated 4 times)
The RPMs installed as reported by yum are:
Interestingly, I don't seem to get these memory errors when I run against
the xccdf directly and not the combined datastream, but then I can't use
the tailoring file I created. Any ideas on the memory errors?