search cancel

Virtual Service Model ended running after 3 errors

book

Article ID: 38127

calendar_today

Updated On:

Products

Service Virtualization

Issue/Introduction

We have noticed that the VS shuts down after 3 consecutive errors. How can we change this behavior?


We would like to have the Virtual service restart automatically. Is there a way to do this?

Environment

All supported DevTest releases and plaforms.

 

Cause

By default, a Virtual Service will go offline after one error.

Resolution

By default, the VSE will shut down after 3 consecutive errors are seen. There's an undocumented property, which can be added to the local.properties file, to change this behavior.

The 'lisa.vse.max.hard.errors' property allows you to configure the number of consecutive errors allowed before the VSE service shuts itself down. 

To modify the default behavior, open the local.properties file on the VSE server, and add the property:

lisa.vse.max.hard.errors=<number>

Where <number> represents the number of allowed consecutive errors. The value can be 0 or greater.

    • A value of 0 means that no errors are allowed, and the service shuts down after the first error is seen.
    • To designate an unlimited number of errors, enter any negative number (-1).( This is NOT RECOMENDED)

For example: lisa.vse.max.hard.errors=100


After adding this property to your local.properties file on the VSE server, you will need to restart your server.

NOTE:
When setting lisa.vse.max.hard.errors to 100 or more you will also need to set the following:
lisa.coord.failure.list.size=10

If the above is not set you can run into Java heap Memory issues and the VS will through 1000s of errors...

Additional Information

Use the monitor feature to enable Auto restart:
Monitor Virtual Services (broadcom.com)

What happens once the number of Errors thrown by a VS exceeds the Max VSE Hard Errors limit?

Answer: That depends on whether the VS was set with AUTO-RESTART or not.

SOME BASICS ON HOW THE AUTO-RESTART FLAG WORKS in Conjunction with MAX VSE Hard Errors Setting

AUTO Restart FLAG: OFF

 1. If the VS is stopped manually from Portal or through any other means, then it will stop immediately and cease listening to incoming requests/messages

2. If the VS is stopped due to some internal errors, then it will immediately stop listening to incoming requests/messages

 AUTO Restart FLAG: ON

 1. If the VS is stopped manually from Portal or through any other means, then it will stop immediately and cease listening to incoming requests/messages

2. If the VS is stopped due to some internal errors, then it will restart on its own till it reaches the limit set by the flag "lisa.vse.max.hard.errors". If not set, the default value for this property is 3. That means, the VS will try 3 times to start itself once it encounters an error

DEMONSTRATION

lisa.vse.max.hard.errors=1000

See the impact of AUTO-RESTART FLAG on VS DE497983 and "DE497983 2". These are 2 identical VS that use IBM MQ. 

DE497983 - Disabled 

DE497983 2 - Enabled

As you can see below.;

First Screenshot - Both VS faced 1 error each

Second Screenshot - DE497983 Stopped because Auto-Restart was disabled; whereas "DE497983 2" continued in spite of the errors and it will survive for the next 998 error (because lisa.vse.max.hard.errors=1000)

See the last screenshot also, you could see that several VS that are set with 'Auto Restart' ON survive errors as the number of errors is well within the 1000 limit.

 

As you can see the extreme case also, the VS comes to a stop once the number of errors cross the MAX VSE HARD ERRORS Limit of 1000 set.

Attachments