02-28-2014 04:39 AM
Hello,
are any circumstances known where a remote engine might be restarted automatically?
We are searching for sporadic problems in a remote execution or in the communication with the remote engine.
Several remote executions are started during ProcessSetup on the controlling TestStand instance, all using the same host name, so only a single remote engine is started on the remote computer. The remote executions run in an infinite loop which is stopped during ProcessCleanup on the controlling TestStand instance by a command sent through a queue.
Yesterday I noticed that on the remote computer two remote engines were running. One was "old", i.e. started two days ago, one was new, started only a few hours ago. The "old one" apparently still had the queue connections.
The only method I know of to start several remote engines on the same computer is by starting them with different host names, i.e. one instance using the computer name, the other instance using the IP address.
We also see -17500, lost connection to synchronization manager, errors sometimes on this machine, as well as -17300, queue does not exist, errors.
Is it possible that a temporary loss of network connectivity causes an additional remote engine to be started?
Best regards
Peter
02-28-2014 09:13 AM
It sounds plausible that this might be caused by a lost connection. Ultimately we are using the Windows DCOM functionality underneath so however that behaves is what you will get. You might try searching on google or MSDN to see how DCOM handles such a situation.
-Doug
04-03-2014 01:59 AM
Actually, it was a trivial reason, nevertheless I do not understand the behavior.
The reason was a wrong configuration of the power supplies, leading to the controlling PC being restarted while the remote PC remained running. Autostart on the controlling PC then restarted the main TestStand sequence which in turn started the sequences in the remote engine.
But what I do not understand: when I make several remote sequence calls with the same combination of host and remote addresses / computer names, I am used to the same REngine being re-used. Why does a second (and sometimes, 3rd and 4th, depending on the number of restarts) instance get created in this case?
Regards
Peter
04-03-2014 02:15 AM
Peter,
does the Main PC have a static IP?
I find it possible that during the force-reboot, the PC receives a new IP. Which should create a new REngine instance on the "satellite" as the already running instance cannot hook up due to a different IP requesting the REngine....
Norbert
04-03-2014 03:28 AM
The IPs are fixed which is typically the case in our production environments.
Thank you
Peter
04-07-2014 04:15 AM
Peter,
does the REngine always start as new process? So if you reboot the master on purpose during execution, is it behaving the same every time?
Or does it, in certain situations, pickup the previous instance?
Norbert