FORMFACTOR key is Unknown instead of desktop. This is most likely of no value and is probably wrong in most cases, so if it's not easy to fix we should just drop the key.
One value that *is* useful to have is FORMFACTOR="blade" as in the IBM JS20 systems. So we should make sure this is fixed.
We do not have a FORMFACTOR field in Beaker's DB..so we will need to fix that as well.
beaker-system-scan patch: http://gerrit.beaker-project.org/#/c/4257/ lshw doesn't report the formfactor on the IBM js20 system correctly, so need to patch it..
(In reply to Amit Saha from comment #4) > > lshw doesn't report the formfactor on the IBM js20 system correctly, so need > to patch it.. lshw uses SMBIOS data for the "description" of the computer which is equivalent to what we need for "form factor". However, SMBIOS doesn't exist for PPC (http://www.dmtf.org/standards/smbios). I am looking into hal's source code to see if I can understand how smolt is able to get it. The file specifically is hald/linux/osspec.c (searching for "formfactor" will lead to the key places in the code).
We adopted a strategy similar to smolt specifically for IBM blade systems which is use the model to check if it's a Blade, then we set it so in beaker-system-scan. We are going to work on a patch for lshw to do so, but in the meanwhile we strictly do the mapping in beaker-system-scan: http://gerrit.beaker-project.org/#/c/4257/
(In reply to Amit Saha from comment #6) > We adopted a strategy similar to smolt specifically for IBM blade systems > which is use the model to check if it's a Blade, then we set it so in > beaker-system-scan. We are going to work on a patch for lshw to do so, but > in the meanwhile we strictly do the mapping in beaker-system-scan: > > http://gerrit.beaker-project.org/#/c/4257/ We have patched/plan to patch lshw to report the chassis correctly, hence beaker-system-scan just retrieves it from lshw output instead of doing any mapping itself.
Beaker 21.0 has been released.