dpol1 commented on issue #9016:
URL: https://github.com/apache/storm/issues/9016#issuecomment-5694293440

   I will give it a try, then.
   
   If I read your model right, the tuple arriving in `status` already carries 
`0x0003`, but nothing reports it. A bolt's `fail()` sends only root ids to the 
acker, and the spout sees a timeout only for the root. How would you tie the 
timeout to the branch that carried the label, especially after fan-out, where 
many tuples share a route?
   
   Also, a stream can have several subscribers (in StormCrawler, WARC and the 
parser both read `fetch`), so `(prefix, nextStreamOut)` is ambiguous at emit. 
Swapping on arrival, keyed on the receiving bolt, avoids that.
   
   My use case needs the HTTP and parsing work tied to one fetch attempt, so I 
will first check that the context survives the fetcher's threads, with 
sampling. Labels can come later, once multiple anchors and the table size have 
an answer. I'll write up the design and the test cases here.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to