Description of problem: currently HTS will treat systems which "Model" or "Vendor" information is uncertain(can not be found by HTS) as PV guests. and record this PV info in the result rpm names. the related code is in hts/hts/hardwaretest.py : 634 if not (foundModel or foundVendor) and os.system("hal-device | fgrep -q xen") == 0: 635 if self.Debugging: print "Running in a PV Guest" 636 self.certification.setHardware(Tags.vendor, "Red Hat") 637 foundVendor = True 638 self.certification.setHardware(Tags.model, "ParaVirt Guest") 639 foundModel = True But the machines which "Model" and "Vendor" are unknown to HTS do exist, take this one as an example: https://hardware.redhat.com/show.cgi?id=460990 https://hardware.redhat.com/hardware.cgi?cert_id=460990&id=140825 https://hardware.redhat.com/results.cgi?cert_id=460990&id=157267 this is definitely not a PV guest.
The current logic is only applied when there are hal devices with the string "xen" in them.
Isn't this already open a different bug to fix the PV vs. FV vs. Dom0 vs. BM which yk is already working on?
bug #310861 I believe?
(In reply to comment #1) > The current logic is only applied when there are hal devices with the string > "xen" in them. I guess this is not enough to determine a system is PV ?
(In reply to comment #3) > bug #310861 I believe? this should be another issue, about the result rpm name. HTS will include both Vendor and Model information into the result rpm names. On the systems which has the string "xen" in the hal devices, if either "Model" or "Vendor" information is not available to HTS, HTS will treat it as PV. this is not correct because such machines that have the string "xen" in the hal devices when running on the "xen kernel" do exist, e.g.: https://hardware.redhat.com/show.cgi?id=460990
YK, do you have patch for this that you'd like to propose?
I'd suggest to simply remove this part of code from hts/hts/hardwaretest.py . there's no need to check if the system is PV here(bug#310861 will do that check). The purpose of "getBiosInfo()" is just to gather information for packaging the result rpm. it should proceed to "ask humans for it" if it can not find the machine model and vendor by it's own. Do you agree with my suggestion ? If yes, then I can create the patch immediately. -YK
If Greg'll sign off on both patches works for me.
Sounds good to me.
patch submitted, removed the PV checking here, please review. Thanks -YK
Looks good to me.
Verified in hts-5.3-12 hts would ask user input if the "Model" or "Vendor" was unknown and checked unknown field in result.xml file of "Model" or "Vendor".
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2009-0047.html