[erlang-questions] Best practices for distributing an application over erlang?
Fri Mar 11 23:12:49 CET 2011
On 2/17/11 1:53 PM, Scott Lystig Fritchie wrote:
> Gregory Haskins <> wrote:
> gh> Long story short(er): I suspect I am just making all of this harder
> gh> than it needs to be, and there must be a ton of you who are just
> gh> doing similar operations all the time.
> Er, it seems to me that you are mixing together two problems:
> 1. managing application bugfixes & upgrades
> 2. managing distributed application state
> Having a whiz-bang way to manage the BEAM files required to do #1 is
> usually done by copying the new BEAM files to all the places that they
> need to be, then using a. load a module at a time (usually via the
> console, for small patches), b. OTP app upgrade (cool but tricky to
> coordinate, depending on the app), or c. rolling node stop & restart.
> There are other ways to do it, as you were exploring, but they aren't
> the sort of things that I'm aware of many (any?) people using in
> production. (Aside from diskless/stateless embedded systems.) Or I'm
> completely misunderstanding your question....
Thanks for the detailed reply, Scott. I was afraid of your answer, but
I started putting together a prototype to address this apparent
shortcoming. I call it "EDist", short for erlang-distribution. If
anyone is curious, I have the code up on github:
Ill update the wiki with better notes on what it does today and what I
want it to do in the future. Note that it is still very much in a
prototype form, and I fully expect that the protocol ABI will change
before I get to "1.0".
Patches welcome, of course.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 267 bytes
Desc: OpenPGP digital signature
More information about the erlang-questions