[erlang-questions] Problem attaching a shell to an Erlang node started through SSH
Fri Jul 6 07:27:25 CEST 2012
This is what we're doing (see my recent email) and the erlang is working fine. I'll be at the office this afternoon to check properly but here are a few differences:
- our fab script uses sudo() instead of run();
- this function calls an unit script instead of calling the release start directly.
> However, attaching a shell to the
> node triggers a strange
> behavior, the shell is available but
What are the symptoms?
Sent from my iPhone
On 5 Jul 2012, at 23:27, André Graf <> wrote:
> I am working on a small library that uses Python Fabric together with
> Rebar deploying of Erlang applications. Fabric provides a scriptable
> (it's Python) abstraction over SSH.
> While I am able to deploy and upgrade new releases through such a
> fabric script I struggle with a simple Erlang node startup.
> A fabric call e.g. run("/path/to/my/release/bin/myproject start")
> invokes following SSH command
> ssh "/bin/bash -l -c '/path/to/my/release/bin/myproject start'"
> This starts the node and I can use the remote shell to operate on the
> node. However, attaching a shell to the node triggers a strange
> behavior, the shell is available but unusable. It seems the startup
> through SSH has somehow confused the involved pipes.
> Do you have any hints on that? I am using rebar 2.0.0 R15B
> 20120625_185716 git 2.0.0-45-gc41fda6.
> erlang-questions mailing list
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the erlang-questions