On crash or stop of the image or the watchdog, the RabbitMq queue should be voided and emptied

tesonep / pharo-testWorkers

4 stars 4 forks source link

On crash or stop of the image or the watchdog, the RabbitMq queue should be voided and emptied #5

Closed VincentBlondeau closed 11 months ago

VincentBlondeau commented 3 years ago

In a kill event of the images (master or slave), it seems that the queues of the rabbit mq are not emptied and can still contain some requests: run some tests that are in the queue buffer (not delivered yet) or pong answers. I think that it is due to the fact that the same queue is used again to interact between the master and the workers.

There should be a process to either :

clean the queue
create a new queue with a random id

tesonep commented 3 years ago

@VincentBlondeau I have fixed this, can you try it?

fedemennite commented 11 months ago

This was indeed working for us since quite some time, Can be closed at your convenience