Web REST service shutting down

[dimwits86]

Hi,

I'm experiencing some strange issues with the Web REST service (running as an application) from which I'm using the JSON data to run my own status monitor webpage. Unexpectedly and apparently without any reason it just stops running. I turn it on, it works for a few hours and then it shuts down, having to restart it via the config window. Jobs are running as normal.

Also, there seems to be an issue with the monitor cache - it doesn't clean up properly, as seen in the picture attached. Those are jobs that completed successfully about a day ago but the entry still persists in the list (only on the web monitor/JSON data, not in the app). It can be fixed by deleting the files in /Processors/.cache/monitor/jobs but if I set a script to do it periodically would it not kill jobs that were currently running at that time?

[admin]

Hi dimwits86,

What version of the rest_service.exe are you running? Check the about-section in the FFAStrans manager gui.

-steinar

[dimwits86]

Hi steinar,

I'm on 0.9.3.13. Is there a newer one?

[admin]

Actually, I just had the same thing happening on my Win8.1 test PC...which is a first-time. Currently I'm not sure what caused it but kind of surprised me. So I need to figure out what might have caused it. At work we're running 6 servers and this has never happen. Also, I've fixed some code that should prevent FFAStrans from creating "dead/fileless" monitoring-records.

-steinar

[dimwits86]

So could it be an OS issue then? I'm running a farm with four Windows 7 nodes and it seems to happen on all of them. The process just goes down after a few seconds or sometimes minutes/hours. I can probably work out a script of some sort to re-launch the process if it closes.

[admin]

Currently, I don't know why it happens but I have discovered a weakness in the FFAStrans manager when running as application. The manager should restart the API automatically if it goes down. This does not happen so I will fix that. There should not be any need to script your way out of the problem. However, the best thing would be if the API did not go down, and this I'm currently investigating.

-steinar