Hi Suho, As per the offline discussion with @Shehan Panditharatne <[email protected]>, the aggregation was run independently, however with the same data source. The issue we faced was that when in-memory buckets were calculated from the database, there were overlaps. For instance when "SP1" node has already stored the second data, then only the SP2 node gets the first message, it had loaded "SP1" nodes aggregation from the database, thus we faced duplicate key exceptions.
AFAIK, this shouldn't occur if we use different data centres. Best Regards, *Niveathika Rajendran,* *Senior Software Engineer.* *Mobile : +94 077 903 7536* On Sun, Nov 4, 2018 at 5:01 PM Sriskandarajah Suhothayan <[email protected]> wrote: > Hi Nivedika > > This seems not to be possible, can you explain reason for failure? > > Regards > Suho > > On Sat, Nov 3, 2018 at 5:56 PM Shehan Panditharatne <[email protected]> > wrote: > >> Hi, >> >> I have attached the modified siddhi app for two parallel worker nodes. >> I have added the attribute node to PreProcessedESBStatStream and >> PreProcessedMediatorStatStream. It will be "sp1" in the first instance and >> "sp2" in thee second instance. >> >> Best regards >> >> -- >> Shehan Panditharatne >> Software Engineer Intern | WSO2, Inc >> Mobile: +94771554990 >> > > > -- > *S. Suhothayan* | Director | WSO2 Inc. <https://wso2.com/> > (m) (+94) 779 756 757 | (e) [email protected] | (t) @suhothayan > <https://twitter.com/suhothayan> > GET INTEGRATION AGILE > Integration Agility for Digitally Driven Business >
_______________________________________________ Dev mailing list [email protected] http://wso2.org/cgi-bin/mailman/listinfo/dev
