uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eddie Epstein <eaepst...@gmail.com>
Subject Re: Uima Ducc Service restart on timeout
Date Sat, 29 Oct 2016 15:14:15 GMT
Hi Wahed,

One approach would be to configure the service itself to self-destruct if
processing exceeds a processing threshold. UIMA-AS error configuration does
support timeouts for remote delegates, but not for in-process delegates. So
this would require starting a timer thread in the annotator that would call
exit() if not reset at the end of the process() method. DUCC will
automatically attempt to restart a service instance that exits.

DUCC's pinger API allows a service pinger to detect that a service instance
is not working correctly and tell DUCC to restart the instance. This
approach has been confirmed to work for the current trunk code, after post
v2.1.0 fixes.

Eddie


On Fri, Oct 28, 2016 at 10:07 AM, Wahed Hemati <hemati@em.uni-frankfurt.de>
wrote:

> Hi,
>
> is there a mechanism in Ducc to restart a service, if it is processing a
> cas for to long?
>
> I have a annotator running as a primitive service on Ducc, which sometimes
> gets to an endless loop. I call this service with a Uima AS Client. I can
> set a timeout on the UIMA AS Client, this works. The client throws a
> timeout after the specified period, which is nice. However the Service is
> still processing the cas somehow? Can i tell Ducc to shutdown and restart a
> service, if it is processing a cas takes more then a specified time period
> (e.g 60 sec)?
>
> Thanks in advance
>
> -Wahed
>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message