0

Title should be pretty self-explanatory.

I'm specifically interested in Cascalog, but I might accept an answer tuned more broadly to Cascading if it seems clear how that might apply towards Cascalog.

Occasionally, I'll create a Cascalog query that does the wrong thing and attempts to iterate through some massive collection on my little development laptop. When this happens, I presently have no way of killing the job and salvaging my precious battery life sans killing the parent Clojure process, which always bites (never fun rebooting the JVM/Clojure and getting app state back again...).

metasoarous
  • 2,854
  • 1
  • 23
  • 24
  • How do you know when does a job go "wrong"? Does an independent process notify you? Or your Cascalog application is aware of it is "wrong"? – chinglun Dec 12 '15 at 01:02
  • I'm not asking for automated "gone wrong" detection. If I realize something has gone wrong while developing, I'd like to be able to send a kill message. – metasoarous Dec 13 '15 at 04:39
  • Talking about a manual kill on a local machine, this post might help: http://stackoverflow.com/questions/11458519/how-to-kill-hadoop-jobs – chinglun Dec 14 '15 at 23:44

0 Answers0