[erlang-questions] receiving pang from net_adm:ping/1 even though net_adm:names/0 shows the other node

Dennis Byrne <>
Sun Apr 19 19:52:57 CEST 2009


On Sun, Apr 19, 2009 at 12:43 PM, Per Melin <> wrote:
> On Sun, Apr 19, 2009 at 7:30 PM, Dennis Byrne <> wrote:
>>> Try starting both nodes with -sname. That doesn't require a valid hostname.
>>
>> This is what I did in the original post, starting the nodes with $ erl
>> -sname foo -setcookie c and $ erl -sname bar -setcookie c .
>
> Well, if you go back and look what you posted that wasn't actually it.

You are correct.  I looked back and it looks like the bar node was
brought up with -name instead of -sname .   I think this was a typo
however because I have reverified same problem using -sname.  Also,
-name wouldn't have given the next console line of "" - that is
how erlang behaves with -sname (as you probably already know).  Thanks
for noticing this though.

Is there any way to force erlang to just use an IP address rather than
do a DNS lookup?  I always find myself wanting to do this when I debug
issues like this.

> Anyway, I tried to change my hostname to 207 and then the same thing
> happened to me as you describe in your original post. It seems it is a
> bad idea in general to have a numeric hostname, since there is no way
> to tell it apart from an IP-number.
>
> fanboy:~$ sudo hostname 207
> fanboy:~$ ping 207
> PING 207 (0.0.0.207): 56 data bytes
>



-- 
Dennis Byrne



More information about the erlang-questions mailing list