[erlang-questions] Ideas on Distributed Programming on single machine

lloyd <>
Sat Jan 17 21:23:16 CET 2015


Hi Mark,
 
Are you by chance running riak or riak CS on any of your machines?
 
Thanks,
 
LRP
 
-----Original Message-----
From: "Mark Nijhof" <>
Sent: Saturday, January 17, 2015 2:53pm
To: "Christopher Meiklejohn" <>
Cc: "erlang questions" <>
Subject: Re: [erlang-questions] Ideas on Distributed Programming on single machine



I have had 14 different machines running on my 2 year old MacBook Air (granted they where not doing a whole lot) by just using docker. Each docker instance has its own IP and name ect. Worked really well. Make sure that in each docker you run tmux so yo can check both the output and do things with the machine. 
What I do is put the Dockerfile in a sub folder of the project (because docker copies all things in the folder the Dockerfile is hosted in into the container and that is slow) and then map ../ to a mountpoint inside the docker container. If your host is the same as the docker container then you can build locally and just restart the process in the containers.

-Mark


On Sat, Jan 17, 2015 at 8:46 PM, Christopher Meiklejohn <[  ]( mailto: )> wrote:




On Jan 17, 2015, at 8:23 PM, Harit Himanshu <[  ]( mailto: )> wrote:




This is really dumb question and I am pretty sure that there is limit to how much distributed programming could be learnt using single machine(I am using Mac in this case).



I am reading through Programming Erlang, Chapter 14, Distributed Programming where Joe talk about how to run Name Server in distributed mode as 


 I write and test my program in a regular nondistributed Erlang session. This is what we’ve been doing up to now, so it presents no new challenges.
 I test my program on two different Erlang nodes running on the same computer.
 I test my program on two different Erlang nodes running on two physically separated computers either in the same local area network or anywhere on the Internet.
You can get pretty far using all of the networking tools provided with your operating system to simulate various network conditions.  I do all of my daily distributed programming and research using a stock MacBook.

Consider Kyle Kingsbury’s work on Jepsen where he’s able to find bug in several major distributed databases.  There are also tools such as ‘tc’ which can be used to add arbitrary latency between processes.
For what it’s worth, Basho does a significant amount of fault testing on a single machine by using fault-injection tools, or by facilities like ‘intercepts’ provided by Riak Test.
My advice is this: do everything locally on your machine until you have a reason to move to separate infrastructure; it will be easier to develop and debug.
- Chris


Christopher Meiklejohn
Senior Software Engineer
Basho Technologies, Inc.
[  ]( mailto: )
_______________________________________________
 erlang-questions mailing list
[  ]( mailto: )
[ http://erlang.org/mailman/listinfo/erlang-questions ]( http://erlang.org/mailman/listinfo/erlang-questions )



-- 


Mark Nijhof


t:   [ @MarkNijhof ]( https://twitter.com/MarkNijhof )
s:  marknijhof
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20150117/56f19ce0/attachment.html>


More information about the erlang-questions mailing list