DL 2.7.0.7
win 7 x64
Looks like when the Balancer runs in background / service, it isn’t working right. As we suspected. Log sample below. At 8:34:57 where it should say “requesting 20 new instances be started” it says "Object reference not set to an instance of an object.”
After I stop the service and launch the balancer in the foreground as a GUI on the same node (ZA01), it works as expected. See 8:37:25 entries.
Let’s take this to Thinkbox along with the question about the 32-core GCE machines?
2015-07-23 08:34:57: Group: cloudgroup Target: 20
2015-07-23 08:34:57: Algorithm (16): Targets:
2015-07-23 08:34:57: Algorithm (16): Region: us-central1-f
2015-07-23 08:34:57: Algorithm (16): Group: cloudgroup: 20
2015-07-23 08:34:57: Worker Thread (16): Equalizing targets…
2015-07-23 08:34:57: Object reference not set to an instance of an object.
2015-07-23 08:34:57: Worker Thread (16): Waiting 60 seconds before next cycle.
I’ve asked around, and we don’t have any immediate ideas on what might be causing the “value cannot be null” issue. If this is preventing Balancer from working for you, we can schedule a remote session so that we can look at your configuration and watch the behavior first hand. Let us know if you would like to do that.
the “value cannot be null” is not stopping us. Just an error that looks like something somewhere isn’t right, so since we where under the hood, thought we’d bring it up. and I’m actually referring to we in the ‘our engineering dept’, not the royal We…that’s only on we-ekends.
We should have a developer looking at the “Balancer as a Service” issue yet this week or early next week. However, I would not expect any news until after Siggraph next week.