Fork me on GitHub
#onyx
<
2016-02-21
>
robert-stuttaford08:02:40

@lucasbradstreet, sometimes when i submit jobs, i get lots of the following, which prevents the start-job process from exiting

robert-stuttaford08:02:09

which prevents the peer daemon process from starting

robert-stuttaford08:02:27

the actual start-then-kill-jobs stuff appears to run fine

lucasbradstreet08:02:35

Hmm. Are you submitting the job before the peers have started? I thought that works but maybe I was wrong

robert-stuttaford08:02:47

submit, kill, stop, start

lucasbradstreet08:02:00

Okay, maybe I was wrong about that then. I'll check.

lucasbradstreet08:02:08

I'm not on a computer atm

robert-stuttaford08:02:15

sorry to pester you on a sunday. let’s tackle tomorrow

robert-stuttaford08:02:26

enjoy the rest of your weekend!

lucasbradstreet08:02:35

No problem anyway. Thanks for the report

lucasbradstreet09:02:53

@robert-stuttaford: looks like submit job works OK even when the peers aren’t started up yet, so I’m going to have to ask you some more questions when you’re around

robert-stuttaford15:02:05

ok lucasbradstreet - we can tackle it tomorrow simple_smile

robert-stuttaford18:02:17

hey lucas, are you around?

robert-stuttaford18:02:25

oh, i see you’re Z, no worries

Drew Verlee22:02:58

I'm curious how onyx compares to google's dataflow. They seem to share some underlying principles. I feel i could learn a lot by trying to compare the two. http://blog.acolyer.org/2015/08/18/the-dataflow-model-a-practical-approach-to-balancing-correctness-latency-and-cost-in-massive-scale-unbounded-out-of-order-data-processing/

gardnervickers23:02:14

@drewverlee: Onyx’s state management features are based off of the dataflow paper