facter in Debian now recommends virt-what. facter automatically uses it
over its own internal detection if it available.
virt-what looks like it has more comprehensive coverage over facter. For
example, in an Openstack deployment I tried just now, in Saucy virt-what
correctly detects KVM, whereas "facter is_virtual" returns false unless
virt-what is installed.
I've just prepared a merge for facter, but have had to demote virt-what
from a recommendation to a suggestion as it is universe. With virt-what
in main, we could reduce the Ubuntu delta in facter, and users will pull
virt-what in by default for better detection.
This may make bugs like 1170325 moot. virt-what is just shell and a tiny
bit of C, so SRUs shouldn't be as likely to hit issues like bug 1173265
either.
Caveat: virt-what requires root. But for the common case of facter being
used with puppet, puppet runs as root anyway.
Is it worth considering an MIR for virt-what so that facter works
better?
A quick look:
* Has a test suite.
* No outstanding bugs: neither in Ubuntu nor Debian.
* Red Hat uses it too, and it looks like they are actively maintaining
fixes for it[1]. It looks like they're upstream too, though I'm not
sure.
* Only a dmidecode dependency (already in main).
* I don't think it's security sensitive.
Possible negatives:
* Uses 1.0 packaging so harder to maintain patches.
* Doesn't currently run the test suite during the package build, so we
may need a delta for this if Debian don't accept a patch.
[1]: http://rhn.redhat.com/errata/RHEA-2013-0483.html
--
ubuntu-server mailing list
ubuntu-server@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-server
More info: https://wiki.ubuntu.com/ServerTeam
