storm-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Harshal Vora <harshal...@gmail.com>
Subject Storm DRPC - Exception Handling ?
Date Sat, 02 Jan 2016 05:19:47 GMT
Hi,

I have a trident topology which I'm using to parallelize some heavy 
on-the-fly computation. I'm having trouble handling runtime exceptions 
which cause the worker to cause and restart every time. I would like to 
suppress all exceptions within my functions/filters, but at the same 
time I want the drpc output to reflect that there were errors in the 
computation (by setting a flag somewhere). I could achieve this by 
setting a flag field in the tuple as it passes through the topology, but 
this would require me to create a separate flag field for every 
function/filter, which leads to very dirty code. Another way would be to 
maintain this flag in the topology context using some unique id for each 
drpc invocation, but in all the documentation it doesn't say if such a 
thing is possible and if yes then how to do it.

Any ideas on how to achieve this? I could use the 
TridentCollector.reportError() method to report errors to the UI, but 
this doesn't help me, I need to be able to refer to the errors in the 
topology itself.

Regards,

-- 
Harshal Vora
www.radiolocus.com


Mime
View raw message