[erlang-questions] Releases and Different OSs
Wed May 18 13:06:35 CEST 2016
I have four prebuilt ERTS for centos 32/64 and debian 32/64. ERTS were
configured with --disable-dynamic-ssl-lib --without-termcap
flags and built on centos 5.8 and debian 6 They are still compatible with
the latest versions of Centos/Ubuntu/Debian, which are used on 99% of
commercial hosts. Some library apps which I don't need are stripped as
well. Then ERTS together with my app is packed into deb and rpm packages.
All paths' and environment variables are hardcoded in startup scripts. This
saves me from a lot of headche, as I have a lot of small customers without
admins at all.
If you have inhouse app this would be very different of course.
2016-05-18 3:42 GMT+03:00 Andrew Berman <>:
> Hi List,
> So today I tried to create a release using rebar3 from my Mac and deploy
> it to a CentOS box. Everything works when I do not include ERTS (and have
> Erlang installed on the target), but if I do include ERTS, things fail
> (can't run run_erl), I'm assuming because the Mac version of Erlang can't
> be run on CentOS, which makes sense. So, I wanted to know how people do
> deployments of their Erlang releases. Do you set up boxes with each OS
> that matches your target OS and compile on there, including ERTS in the
> release? Do you not include ERTS and make sure that all your Erlang
> installs are the exact same version? Just trying to figure out a good
> Thanks for any advice!
> erlang-questions mailing list
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the erlang-questions