Disco slave timeouts

[james.m...@infinityworks.com]

Hi,

I thought I had this working at the end of last week but somehow I've introduced a regression and now I can't get the smoke test to run correctly.

I have a master with no workers and 3 slaves with 1 each.

I can ssh from the master to any slave and vice-versa. Also can ssh from any slave to any other. No password prompting.

I have copied the correct .erlang.cookie to every node.

I have run 'healthcheck.sh' on the master and it returns 'pass' for every node.

But the nodes still have a red bar in the UI and they won't accept jobs.

I've tried ssh'ing on to a slave box and executing:

disco@discomaster:~/disco$ disco debug
Erlang R16B03 (erts-5.10.4) [source] [64-bit] [async-threads:10] [kernel-poll:false]

Eshell V5.10.4  (abort with ^G)
(disco_8989_master@discomaster)1> slave:start(discoslave0, "testname").
{error,timeout}

Which implies there's an problem, probably around authentication but I've followed the troubleshooting instructions to no success.

Can anyone suggest some thing to try?

Thanks

[jsha...@gmail.com]

Are you sure all the ports are open?