[ejabberd] Re: Cookie Confusion

Bill Brazell bill.brazell at cingular.com
Wed Apr 20 00:01:51 MSD 2005


OK.  This is starting to blow my mind.  The part that hurts is that I've done
this before.  Now, I can't get the .erlang.cookie file to be written.  When
logged in as ejabberd, I perform the following:

su ~
[ejabberd at jabbertest ~]$ erl -setcookie myCookie
Erlang (BEAM) emulator version 5.4 [source] [hipe]

Eshell V5.4  (abort with ^G)
1> myCookie
1> halt().
** 2: syntax error before: halt **
1> halt().
[ejabberd at jabbertest ~]$ ls -la .erlang.cookie
ls: .erlang.cookie: No such file or directory

Shouldn't the .erlang.cookie file be created at the first command?  Entering the
cookie value within the Erlang console doesn't appear to do anything except
invalidate the subsequent command.

Another attempt:

su ~
[ejabberd at jabbertest ~]$ erl -sname node1 -setcookie myCookie
Erlang (BEAM) emulator version 5.4 [source] [hipe]

Eshell V5.4  (abort with ^G)
(node1 at jabbertest.wireless.attws.com)1> erlang:set_cookie(node1 at jabbertest,
'myNewCookie').
true
(node1 at jabbertest.wireless.attws.com)2> erlang:get_cookie().
myNewCookie
(node1 at jabbertest.wireless.attws.com)3> halt().
[ejabberd at jabbertest ~]$ ls -la .erlang.cookie
ls: .erlang.cookie: No such file or directory
[ejabberd at jabbertest ~]$

Why wasn't the .erlang.cookie file written?  Any clues?  I'm sure it's a simple
solution, or at least I hope so.

I suppose the best method for explaining the cookie creation process would be to
stick with the 1-2-3 method.  I do think the new text is more clear, but only
one method needs to be presented.  Actually, I'm documenting the process as I go
along.  I'll gladly share it once I get if finalized.

BB



More information about the ejabberd mailing list