[Libvirt-cim] [PATCH 0/8] Support for full function consoles

Viktor Mihajlovski mihajlov at linux.vnet.ibm.com
Tue Sep 10 09:35:01 UTC 2013


On 09/09/2013 04:02 PM, John Ferlan wrote:
> 
> Not quite sure where it is, but there is a bug somewhere in patches
> 2->5 as the cimtest 'RASD' for '03_rasd_errs.py' fails to find/add
> the console device. The error message is "Expected 7 RASDs, got 6".
> 
> I also think there's an ordering problem.  If I add the patches 1
> at a time, patch #2 will cause the same cimtest to return 0 (zero)
> RASD's (as found through debugging the cimtest test) if patches
> 3 and 4 are applied.
> 
> I decided to try an experiment to see if I could narrow things down.
> The experiment was add each patch 1 at a time, then add patch 2 to the
> set.
> 
> 1,3 -> Returns found 6, then add patch 2 and returns found 0 RASD's
> 1,3,4 -> Returns found 6, then add patch 2 and returns found 0 RASD's
> 1,3,4,5 -> Returns found 6, then add patch 2 and returns expected 7 found 6
> 
> At the very least patch 2 needs to go after patch 5 or be merged with
> patch 5 since it seems that's where the connection/implementation occurs.
> 
> I didn't dig deep into that test to try and figure out why the console
> device was "missing".
>
Strange, I have not seen this 6/7 message. This means that the test
domain has less RASD types than RASD types found on the system.
Can it be that you have some domain already defined before you run
cimtest? This might explain the behavior (but would be a bug in
cimtest).

As for the patch grouping, it was meant to make the patches more
readable rather than to be technically necessary. In fact, the provider
patches (5,6,7) and the MOF/schemal patch (2) must go together.
I can squash them in a V2, but this could in my opinion wait until
I get more feedback.

Could you do me a favor and apply this here, rerun the RASD test group
and paste the resulting cimtest.log from the RASD directory? Thanks!

-- 
diff --git a/suites/libvirt-cim/cimtest/RASD/03_rasd_errs.py b/suites/libvirt-cim/cimtest/RASD/03_rasd_errs.py
index d582ffb..024b0ae 100644
--- a/suites/libvirt-cim/cimtest/RASD/03_rasd_errs.py
+++ b/suites/libvirt-cim/cimtest/RASD/03_rasd_errs.py
@@ -68,9 +68,11 @@ def init_rasd_list(virt, ip, guest_name):
                 logger.error("Unable to parse InstanceID: %s" % rasd.InstanceID)
                 return rasd_insts, FAIL
 
+            logger.info("RASD found %s for %s", rasd.Classname, guest) 
             if guest == guest_name:
                 rasd_insts[rasd.Classname] = rasd
-
+    
+    logger.info("Expected %d RASDs, got %d", len(rasds), len(rasd_insts)) 
     if len(rasds) != len(rasd_insts):
         logger.error("Expected %d RASDs, got %d", len(rasds), len(rasd_insts)) 
         return rasd_insts, FAIL




Mit freundlichen Grüßen/Kind Regards
   Viktor Mihajlovski

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martina Köderitz
Geschäftsführung: Dirk Wittkopp
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294   




More information about the Libvirt-cim mailing list