[Openvas-discuss] Wanted: Reports of specific distro versions where OpenVAS installs easily ...

Whit Blauvelt whit at transpect.com
Mon Aug 6 18:26:49 CEST 2012

... and just works after taking the steps for a default installation.

Background: I've so far tried OBS and Atomicorp installs on versions of
Ubuntu and CentOS and Fedora that are reported on the OpenVAS website to be
good, with varying degrees of brokenness, and no complete success. I've also
tried BackTrack, which was far from a happy experience (avoid the USB
install - it's worthless; as a VM it's got issues too).

I'm not totally discouraged by this. OpenVAS is obviously richly featured
and well designed in many respects. In an age of VMs, it really shouldn't
need to run well on more than one version of one distro. I'm guessing that
there's at least one distro version of which that's true, where a straight
forward installation, either from packages or building from source, will
result in an OpenVAS VM that can just be used, doing a pretty good first
approximation of a thorough and informative scan.

Now, I know VMs are available, but for ESXi and VirtualBox, not KVM. And in
my corner of the world KVM is viewed as both superior and the standard (by
those not preferring Xen anyway). This isn't the place to argue that. It's
just that setting up a unique VM host to handle a single VM would be

So I'd most welcome any first-hand reports from those who've found the
winning combination of distro and OpenVAS installation method saying

1. Distro and version used
2. Installation method used (packages or source)
3. Recipe followed (if any)
4. Special steps taken beyond recipe (if any)

I know I can just keep plugging away trying likely combinations. Except I've
done enough of that already to project that could take many more hours of
work than just asking. 

Alternately, it looks like the failure of the Atomicorp package/script
install on Fedora 15 is just one of configuration at some level. It doesn't
segfault like the Atomicorp install on CentOS 6.3, and runs much faster,
even if it still is quite blind to the open ports in the range it's pointed
towards. Perhaps you've done a similar installation and then found what
Atommicorp got wrong (not to slight their large contribution, just to hone
it), and have a correction to suggest?



More information about the Openvas-discuss mailing list