CAP Theorem: if all you have is a timeout, everything looks like a partition

Posted by nkeywal on Monday May 25, 2015 @01:53PM
from the blog.thislongrun.com dept.

http://blog.thislongrun.com/2015/05/CAP-theorem-partition-timeout-zookeeper.html

( 3 comments | 52 points | on HN )

Re:CAP Theorem: if all you have is a timeout, everything loo... by zamalek on Monday May 25, 2015 @04:57PM (#9601718)

Given a supervisor script that automatically restarts the process, wouldn't it be simpler for the process to communicate its own death the next time it boots?

[ Parent | on HN ]

Re:CAP Theorem: if all you have is a timeout, everything loo... by marcosdumay on Monday May 25, 2015 @05:08PM (#9601762)

I was think more on the lines of: given a supervisor that'll restart the process on any kind of trouble, isn't the entire optimization redundant anyway?

Anyway, that's a great wake-up call to improve the supervisor.

[ Parent | on HN ]

Re:CAP Theorem: if all you have is a timeout, everything loo... by nkeywal on Monday May 25, 2015 @05:41PM (#9601850)

> isn't the entire optimization redundant anyway? With ZooKeeper the ephemeral znode is linked to a session, and if you restart the process it will have a different session. So you also need to manage the znode deletion explicitly in this case. It's possible of course, I have a little preference for the 'ultimate line in the script' as it does not rely on the restart but both options are likely good.

[ Parent | on HN ]