Hi,

I read the fault tolerance strategy adopted by Storm on the github page (
https://github.com/nathanmarz/storm/wiki/Guaranteeing-message-processing).
It seems that each tuple generated by the Source are attached with a
randomly generated message id. I am wondering why not just assign a
sequence id (e.g., 1, 2, 3, 4, ...) for each tuple, which can make things
much easier?

Thanks in advance for the kind attention to this mail.

Regards,
Yingjun

Reply via email to