Service Virtualization

Hi,

Whenever my MQ server goes down for several reasons like patching, server restarts etc.,  LISA MQ services will stop with error message.. however there is no inbuilt way to start the service when MQ server comes up .. this cause us lot of issues as we don't have control on mq server and those will be taken down several times in a week.. all those time lisa service goes down and we need to manually make it up once server comes back.. 

can you please suggest if there is any better way to resolve this issue.. this is a major blocker for us now a days.. 

devtest version in use - 8.3 

See my answer on this thread: MQ VS will not start automatically when MQ server starts 

The short version:

1. Add this line to local.properties:

lisa.vse.max.hard.errors=-1

2. Modify your VSM to add a step that delays for some amount of time, this is the maximum amount of time that it will take for the VSE service to "reconnect".  Then change the 'Listen' step so its 'If Environment Error' assertion goes to this new delay step instead of 'Abort'.

These set up a process that will "poll" for the MQ server until it comes back.

Thanks Kevin...

1. by doing first suggestion.. looks like it is easy to solve the problem but do we have any other impacts with this solution.. like performance impact , logs filling fast etc., 

2. Did not get exactly.. example if we have delay of 1  hour  and if the mq server did not come up within one hour will it fail  ?  could you please share if you have sample vsm which shows this logic.. thanks

The performance should not be impacted much if you are polling every say 30minutes. As for log files, you can change the log level to WARN - this should reduce the log file size

Kevin.Bowman I have tried the first option you have mentioned.. 

I have applied the lisa.vse.max.hard.errors=-1 in one of our sand box before apply to all other lisa servers.. Looks like it is breaking and the whole VSE is gone down.. My MQ server has restarted last night and it might take a hour to come back.. so VSE has gone down due to out of memory.. 

I guess this is mainly happened due to MQ services in  keep contacting MQ server.. please see the attached logs ..

Changing lisa.vse.max.hard.errors was step 1.  Did you also do step 2?

> 2. Modify your VSM to add a step that delays for some amount of time, this is the maximum amount of time that it will take for the VSE service to "reconnect".  Then change the 'Listen' step so its 'If Environment Error' assertion goes to this new delay step instead of 'Abort'.

No, I have not tried step 2 because 

1. to test this step I need to wait till mq server down (there is no test instance of mq server )

2. Also I have many MQ VSMs so again that needs to be updated on each VSM and extra step which might down grade performance of service when doing performance testing.

Adding in a delay is essential to preventing the unlimited restarts from crashing something.  If you do (1) without (2) then bad things happen, as you have seen.

It's also not a performance hit.  If your MQ server doesn't go down then the extra delay doesn't get executed.  If your MQ server *does* go down then that's the end of your performance test.