<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<title></title>
</head>
<body>
Turned out my problem was with the spread segment. I used the wrong broadcast
address to name the segment. This caused weird problems like spread daemons
going <br>
down with Alarm(EXIT) errors. There was nothing that I saw to diagnose the
problem<br>
<br>
I was using:<br>
Spread_Segment 172.16.65.255:4803 {<br>
security1 172.16.65.109<br>
security2 172.16.65.110<br>
security3 172.16.65.111<br>
security4 172.16.65.112<br>
}<br>
But should have been using:<br>
Spread_Segment 172.16.67.255:4803 {<br>
security1 172.16.65.109<br>
security2 172.16.65.110<br>
security3 172.16.65.111<br>
security4 172.16.65.112<br>
}<br>
<br>
<br>
Jonathan Stanton wrote:<br>
<blockquote type="cite" cite="mid20020621204944.GB28852@cnds.jhu.edu">
<pre wrap="">I would suggest a few things.
First, I would remove the localhost segment from the spread.conf. I'm not
sure it is a problem, but it isn't needed.
Second, in order to test whether the basic spread configuration is
working, I would suggest using the spuser and spflooder programs included
with the spread distribution. The spuser in particular should be easy to
use to check basic functionality. Below is a sample 'successful' session
using spuser:
with a daemon running on port 8701:
[jonathan@ice5 linux]$ ./spuser -s 8701
Spread library version is 4.0.0
User: connected to 8701 with private group #user#ice5
==========
User Menu:
----------
j <group> -- join a group
l <group> -- leave a group
s <group> -- send a message
b <group> -- send a burst of messages
r -- receive a message (stuck)
p -- poll for a message
e -- enable asynchonous read (default)
d -- disable asynchronous read
q -- quit
User> j test
User>
============================
Received REGULAR membership for group test with 1 members, where I am
member 0:
#user#ice5
grp id is -2133009127 1024691918 1
Due to the JOIN of #user#ice5
User> s test
enter message: hi there
User>
============================
received SAFE message from #user#ice5, of type 1, (endian 0) to 1 groups
(9 bytes): hi there
User> l test
User>
============================
received membership message that left group test
--------------------------
If this works, then the basic configuration is working. A slightly better
test is to join the same group using spuser running on two different
machines and see that they receive each others messages when you do the "s
test" command.
If both of these work, then we have isolated the problem to either a
python issue, or the test programs you wrote.
</pre>
<blockquote type="cite">
<pre wrap="">Finally the first time I configured the /etc/spread.conf. I used fully
specified hosts names in it. This caused the spread daemon to display
some wierd characters and seemed to cause it to be unstable.
</pre>
</blockquote>
<pre wrap=""><!---->
There is a fairly short limit on the number of characters that a hostname
can be. The printing of long names causes some minor cosmetic problem but
it shouldn't cause any actual instability. Can you be more specific about
how it was unstable?
Thanks,
Jonathan
</pre>
</blockquote>
<br>
</body>
</html>