Overview

This article will detail potential causes for Service Fabric install errors. 

1. Terminology
2. Troubleshooting Techniques
   • How to verify if there is more than one SF server
3. Errors During Install
4. Errors Post Install
5. Redeploy Service Fabric
   • Remove Job Email Running Twice

Since Service Fabric issues can be quite complex, if this article does not resolve your issue, please submit a support ticket. 

Terminology

Before going into detail about troubleshooting Service Fabric install issues, it is important to understand the install process and some terminology. The Service Fabric application you are installing is a combination of two installers

1. Part 1 of the install process is to install Microsoft's standard Service Fabric installer, referred to as "the cluster".
2. Part 2 of the install process is to install Solver, Inc's applications that sit on top (installed) of the cluster. 

In this article, the acronym SF is used to refer to Service Fabric. 

Troubleshooting Techniques

The following are useful in troubleshooting SF. These are accessible only from the Service Fabric server

How to verify if there is more than one SF server

Check the registry key in the following registry location that the key is pointing to the correct Repository database. You should only see one of the following keys depending on the bitness of the machine.

HKEY_LOCAL_MACHINE\SOFTWARE\OneStop Reporting\Settings\RepositoryConnection

HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\OneStop Reporting\Settings\RepositoryConnection

Then verify the IP/server name referenced in the below keys in dbo.MasterSettings in the Repository db are referencing the correct Service Fabric server.

• Solver.ServiceFabricSchedule
• Solver.ServiceFabricSignalR

Error During Install

The following are common errors during install 

Errors Post Install

Redeploy Service Fabric Application:

Steps:

1. In troubleshooting a Private Host issue, you find that an SF application is not healthy, and rebooting the SF server has not resolved the issue.
2. You need to restart/redeploy/reinstall an SF application

Restarting the Service Fabric Host Service via the Windows Service menu does not a “restart” in the Service Fabric application.

1. Navigate to the service fabric explorer. Localhost:19080 on the customer's service fabric server
2. Expand on Cluster > application> Solver…. > fabric:/SolverFabric.JobDispatcher
   
3. Remember the name “SolverFabric.JobDispatcher”. We will need this later.
4. Click the down arrow and choose Delete Application
   
5. You will be prompted to confirm this action
   

Provide the requested text, which is just the name of the application, and click Delete Application.

6. The delete process will take 5-10 seconds, and you will notice the UI updated.
   1. If it errors out on this step. Consider restarting the SF server and trying again
7. Refreshing the navigation menu on the left-hand side, there are no applications listed under SolverFabric.JobDispatcherType
   
8. Select SolverFabricJobDispatcherType. In the main grid make sure you select Application Type Versions > All and click
   
   
9. In the popup that appears, remove the text “Type” from the end of the text that appears in the textbox
   
   Initial
   
   
   
   Final
   

This is the same name as documented in step 3.

10. Click Create Application. Similar to the above, this process should take 5-10 seconds to complete. The interface will update with the Application once complete.
    

You have successfully redeployed an application. Repeat these steps for each app, where necessary.

Job Email Being Received Twice

Overview

When a job is run, you receive an email that the job is already running even if the job succeeded. You will get more than one if you are under the success and failure notification.

Cause

This occurs because the Solver Application treats the email template as the distinct but non-differentiated values on the SQL server. So, it sends the email twice. This is NOT due to a job being run twice.

Resolution:

Run the following script on the Solver Database. Note that you may have to change this to BI360W:

use SolverDW
delete from email.template where TemplateName = 'JobAlreadyRunning'

INSERT INTO
email.Template (TemplateName, TemplateContent, Culture, [Version])
VALUES
('xJobAlreadyRunning', '<!DOCTYPE html>
<html lang="en" xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta charset="utf-8" />
<style>
html {
padding: 10%;
padding-top: 20px;
font-family: "Segoe UI", Tahoma, Geneva, Verdana, sans-serif;
}
h1 {
font-weight: 600;
color: #333;
padding: 20px;
}
p {
background: orange;
border-radius: 5px;
color: green;
margin: 20px;
padding: 20px;
width: 582px;
}
.job-details {
margin: 20px;
}
td {
background: #eee;
padding: 5px;
}
.header {
background: #2196f3;
color: #fff;
font-weight: bold;
}
.subheader {
color: #063b66;
font-weight: bold;
}
</style>
</head>
<body>
<h1>Hi there,</h1>
<p>The job, "<strong>{{jobname}}</strong>" will not run because an instance from earlier hour is still running.</p>
<br />
<table class="job-details">
<tr>
<td class="header">Job Name</td>
<td>{{jobname}}</td>
</tr>
<tr>
<td class="header">Original Start Time</td>
<td>{{starttime}}</td>
</tr>
<tr>
<td class="header">Trigger Type</td>
<td>{{triggertype}}</td>
</tr>
<tr>
<td class="header">Next Running Schedule</td>
<td>{{nextruntime}} (if current running one finishes)</td>
</tr>
</table>
</body>
</html>', 'en-US', '34a1b3c8-2ca0-41b3-9dd9-072c421fde73')

This script disables the email template from being used.