FW: why complete latency and failure rate is so high of my spout.

Hi There, I’m new to storm. I set up a storm cluster with 3 machines. One machine I use it as master, I start up zookeeper and nimbus server on it, another 2 machines I use them as supervisor. I got a very high rate of failure of my spout and I can not figure out why. Please see the detai

Re: why complete latency and failure rate is so high of my spout.

Hi There, Yesterday, I changed some configurations of storm settings, right now , the spout failure rate dropped to 0. As shown below: Topology stats Window Emitted Transferred Complete latency (ms) Acked Failed 10m 0s

Re: why complete latency and failure rate is so high of my spout.

The topology.max.spout.pending configures how many messages can be un-acked from each spout before it stops sending messages. So in your example, each spout task can have 10 thousand messages waiting to be acked before it throttle itself and stops emitting. Of course if some of those messages are

Re: why complete latency and failure rate is so high of my spout.

esday, July 30, 2014 at 9:44 AM To: user mailto:user@storm.incubator.apache.org>> Subject: Re: why complete latency and failure rate is so high of my spout. The topology.max.spout.pending configures how many messages can be un-acked from each spout before it stops sending messages. So in your example, each s

Re: why complete latency and failure rate is so high of my spout.

he > meter off and let all cars flow into the crowded highway. The better way is > to turn on the meter, let the car wait a little bit and enter the highway > one by one. > > A little bit too long, thanks a lot for your patience to read it. > > From: Nathan Leung > Reply-To:

Re: why complete latency and failure rate is so high of my spout.

July 30, 2014 at 12:09 PM To: user mailto:user@storm.incubator.apache.org>> Subject: Re: why complete latency and failure rate is so high of my spout. Your ack sequence is right, but the timeout is not just the simple sum of these times. Consider a simple topology similar to yours: spout

Re: FW: why complete latency and failure rate is so high of my spout.

Complete latency can be higher if you have a lot of data pending from the spout (your topology.max.spout.pending is not set), however given the low number of tuples I wouldn't suspect this to be the case. What is the layout of your topology? Is it circular in any way? On Wed, Jul 30, 2014 at 12