Graph Engine Availability Group cannot connect to other servers in Cluster #368
-
|
I am getting a TCP Socket Error of 10061 - (Connection Timed out). Here is my trinity.xml config file: I have clean addressability to the machine, as I can ping successfully.
This has been working, and then it just stopped. My guess may be something in my local network. Has anyone else seen this problem? |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
|
I have spent the entire day carefully sifting through the Configuration code and found that I was not using the capabilities required to set up Graph Engine Cluster. I have been using Graph Engine for six years and am still learning how to use it. I will come back here with more documentation, augmenting what we have. We no longer have access to the Graph Engine C# API/SDK; I am trying not to be sore about it, but I am. I've got a project in the books for 2023 to write a developers guide for the Graph Engine and to get funding to hire a team of six to 8 developers and continue developing the Graph Engine - there is so much we can do. |
Beta Was this translation helpful? Give feedback.
-
|
In the documentation, we do have this information regarding how to configure the Graph Engine Server runtime: A Local node configures the settings for the machine that reads the current configuration file. There can be multiple Local nodes. Multiple Local nodes will be merged. If a machine is configured by both a Local node and a Cluster node, the configuration settings in a Local node have higher priority. If specified, they will override the configurations specified in a Cluster node. A Local node can have an optional attribute Template. I found this to be essential for setting up a Graph Engine App Server instance in a GE Availability Cluster. The question I have here is how the Cluster is initialized. What exactly is the process of Cluster setup; does it matter what server comes up first and how are the other remote ge app server instance started? Do you need to start them manually, or are they automatically started? The "Locally" defined server instance can also be part of the "Cluster" definition so that the other instance can connect. Do you need to have a specially coded "Trinity.xml" file on each machine and configure that machine as the local instance? Do I need a special local configuration on each machine for each Graph Engine Module:
|
Beta Was this translation helpful? Give feedback.
-
|
I figured this out by tracing through the code! See #370 |
Beta Was this translation helpful? Give feedback.
In the documentation, we do have this information regarding how to configure the Graph Engine Server runtime:
A Local node configures the settings for the machine that reads the current configuration file. There can be multiple Local nodes. Multiple Local nodes will be merged. If a machine is configured by both a Local node and a Cluster node, the configuration settings in a Local node have higher priority. If specified, they will override the configurations specified in a Cluster node. A Local node can have an optional attribute Template.
I found this to be essential for setting up a Graph Engine App Server instance in a GE Availability Cluster. The question I have here is how the Cluste…