> From my point of view, the test for a certain move already contain the test
> for the solution length. What am I missing?
While for the "non-empty" test of the solution length this is true, the
existence
of a particular move does not imply the "precisely length 1" property.
Therefore,
I propose to only add the following interdiff.
Interdiff [PATCH stable-2.17 04/14] Add tests for handling of dynamic
utilisation
diff --git a/test/hs/shelltests/htools-balancing.test
b/test/hs/shelltests/htools-balancing.test
index f6da27b..95f82c2 100644
--- a/test/hs/shelltests/htools-balancing.test
+++ b/test/hs/shelltests/htools-balancing.test
@@ -149,10 +149,6 @@ diff -u $T/simu-rebal-merged.tiered
$T/simu-rebal.tiered.original
# Heavy CPU load can even push instances on the more
# crowded node
-./test/hs/hbal -t$TESTDATA_DIR/hbal-dyn.data --idle-default --mond --mond-xen
--mond-data=$TESTDATA_DIR/dyn1.json
->>>/Solution length=[1-9]/
->>>=0
-
./test/hs/hbal -t$TESTDATA_DIR/hbal-dyn.data --idle-default --mond --mond-xen
--mond-data=$TESTDATA_DIR/dyn1.json -C
>>>/gnt-instance migrate -f -n node-01-000 inst-1./
>>>=0
--
Klaus Aehlig
Google Germany GmbH, Dienerstr. 12, 80331 Muenchen
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Geschaeftsfuehrer: Matthew Scott Sucherman, Paul Terence Manicle
