Re: Storm latency with fail() | acker tasks die

What version of storm are you using?  In older versions of storm the timeout check was done once every topology.message.timeout.secs.  So that means nothing will timeout sooner than topology.message.timeout.secs, but could in the worst case be almost 2x that. If I remember correctly that in

Re: Storm access permission best practices

This is all off the top of my head so it might not be 100% accurate, meaning I may have missed something in here. The location that you installed storm can be world readable, there is nothing secret here, but you should not allow anyone to write (simply to avoid unfortunate accidents when you

Re: Storm latency with fail() | acker tasks die

I think the spout still checks once every topology.message.timeout.secs, which means a tuple will time out between topology.message.timeout.secs and 2*topology.message.timeout.secs after being emitted. The spout times out tuples by putting emitted message ids in a rotating map with 2 buckets. A

Re: RE: is stateful bolts production ready?

In addition to the query, what is the intent of stateful bolt since we can just hold state in bolt instance? wangchunc...@outlook.com From: Wijekoon, Manusha Date: 2017-08-10 18:56 To:

Performance optimize on local mode.

Hi,We are replaying our data on Local mode storm cluster. Using jvisualvm, we find most CPU are used by following methods (about 60% CPU

Re: Storm latency with fail() | acker tasks die

Hi, Which newer version do you mean? I am working with storm version 1.1.0 Thanks, Preethini On Thu, Aug 10, 2017 at 3:36 PM, Bobby Evans wrote: > What version of storm are you using? In older versions of storm the > timeout check was done once every

Re: Performance optimize on local mode.

Local mode is quite different from how distributed mode would work. I would suggest that you create a distributed setup(with even a few nodes is enough) - Running a Multi-Node Storm Cluster - Michael G. Noll -  and then profile the individual workers to find what is consuming the most CPU

Re: Regarding Issue with Multi-Lang in Storm While Running Python (WordCountTopology)

Hi Stig, Thank you very much for your help. I upgraded my version to 1.0.4 and it is working fine. I appreciate your help a lot. Regards, Sitakanta Mishra On Tue, Aug 8, 2017 at 12:51 PM, Stig Rohde Døssing wrote: > Hi Sitakant. > > You are right, this doesn't work in local

Re: Regarding Issue with Multi-Lang in Storm While Running Python (WordCountTopology)

Glad it's working for you :) 2017-08-10 20:10 GMT+02:00 Sitakant Mishra : > Hi Stig, > > Thank you very much for your help. I upgraded my version to 1.0.4 and it > is working fine. I appreciate your help a lot. > > Regards, > Sitakanta Mishra > > On Tue, Aug 8, 2017

RE: is stateful bolts production ready?

In our case we prefer to use our own state implementation. After going through the code and reading documentation, following is how I understand it. Could you please see if my understanding is correct? 1. Derive from State and provide an implementation. In the commit (txID) method are we

Storm latency with fail() | acker tasks die

Hi, I have a situation where the bolts ack, but the acker tasks fail (which is expected as per my logic). I am measuring the latency of the topology using timestamps in ack() and fail() methods. *Observations:* *-* *ack() - latency ~ 100ms* *fail() - latency