Increase number of devices for aluminium NETCONF scale test

Do not waste cycles on small device count, let assume we have
1hr. max per run, then we can fit 11-12 runs into 12hr budget,
starting at 1k and increment by 500 and stop at 4k devices
should give us enought time to complete 7 test runs (1k, 1.5k,
2k, 2.5k, 3k, 3.5k, 4k).

Currently we're spending approximately tens of minutes per run:
2200 .. 09:49:35.932 - 10:06:42.953 ~ 17 min.
2400 .. 10:58:29.420 - 11:16:41.775 ~ 18 min.
2600 .. 12:28:04.384 - 13:49:35.439 ~ 22min.
2800 .. 14:15:24.558 - 14:46:23.139 ~ 31 min.

Change-Id: Iee9c46447440eef8f78e55abc36be5e3d839fc6d
Signed-off-by: Miroslav Miklus <miroslav.miklus@pantheon.tech>

diff --git a/jjb/netconf/netconf-scale.yaml b/jjb/netconf/netconf-scale.yaml
index b19841fb42d5766a133bb34df4f75ef7788e1737..46efd0a09d277b009012aec4aac18239da08f700 100644 (file)
--- a/jjb/netconf/netconf-scale.yaml
+++ b/jjb/netconf/netconf-scale.yaml
@@ -63,8 +63,8 @@
           branch: "stable/aluminium"
           jre: "openjdk11"
           robot-options:
-            "-v USE_NETCONF_CONNECTOR:False -v SCHEMA_MODEL:juniper -v INIT_DEVICE_COUNT:400
-            -v MAX_DEVICE_COUNT:3000 -v DEVICE_INCREMENT:200 -v TESTTOOL_BOOT_TIMEOUT:600s
+            "-v USE_NETCONF_CONNECTOR:False -v SCHEMA_MODEL:juniper -v INIT_DEVICE_COUNT:1000
+            -v MAX_DEVICE_COUNT:4000 -v DEVICE_INCREMENT:500 -v TESTTOOL_BOOT_TIMEOUT:600s
             -v MAX_HEAP:4G --flattenkeywords name:*repeat*keyword*"
       - magnesium:
           branch: "stable/magnesium"