That's definitely not what should be happening. Cancelling a running job should not have any effect on the rest of the queue (there is a regular engine reboot after an x number of jobs if i recall correctly, but that too should not have any effect on the queued jobs).

This sounds like something you should send in to support, either Safe or your local reseller, to take a look at.

I completely agree with @Hans van der Maarel​ , this is unexpected behavior.

You could start by verifying the value of MAX_TRANSACTION_RESULT_FAILURES in fmeServerConfig.txt, it should be set to 10 (default). Any other value could indicate that someone has seen a need to modify this value, which could indicate a related problem.

@Hans van der Maarel​ and @david_r​ , thank you both for the reply. :)

so cancelling running job won't cause the engine restarting? I would check with our architect to get more information.

could I please ask if cancelling job won't trigger engine restarting, does the running job expiry time parameter can be used to control the time out for each job? if there any method we could re-run timed out job automatically in certain period?

Again, thanks heaps

Cancelling a job shouldn't cause an engine restart (unless FME Server was planning to do that already) and it certainly shouldn't cause queued jobs to disappear.

You can use the Running Job Expiry Time to automatically cancel a job that runs too long, but that won't resubmit it (because it would most likely take just as much time and then cause it to be cancelled again, putting you in an infinite loop).

this does sound odd for sure - It would be intersing to know what kind of FME Server set up you're working with and which version of FME Server your running.

If you are running on a system with scaling engines it might be that the engine running is a temporary one and (maybe) also it's queue.

Typically if a job is in a queue with no engine then it just gets stuck but maybe there is some kind of cleanup job that kills any stranded jobs?

It doesn't explain the engine restart but perhaps the MAX_TRANSACTION_RESULT_FAILURES is set to 1 as @david_r​ suggested.

Either way it's not normal behaviour and you should be able to cancel a running job without effecting other jobs in the queue.

@Hans van der Maarel​  and @virtualcitymatt​ , thanks heaps for your reply. we found that fmeServerConfig.txt and MAX_TRANSACTION_RESULT_FAILURES is 10 as default. but when i am testing time expiry cancelling, it still trying to restart the engine. like this. we are using high availability architecture, engine VM and core VM are separated VM. would think this is the issue?