[Server-devel] XS_0_4 install]

[Tony Anderson]

The problem with ejabberd turned out to be documented earlier by Bryan. 
A slightly simpler solution works:

Ejabberd fails when the hostname has been changed. The symptom is that:

service ejabberd status

returns ejabberd not running.

Try the following to fix ejabberd:
service ejabberd stop
killall epmd
rm -rf /var/lib/ejabberd/*
service ejabberd start

Tony

-------- Original Message --------
Subject: Re: XS_0_4 install
Date: Thu, 25 Sep 2008 17:35:06 +0545
From: Tony Anderson <tony_anderson at usa.net>
To: Martin Langhoff <martin.langhoff at gmail.com>,  Bryan Berry 
<bryan at olenepal.org>, server-devel at lists.laptop.org
References: <48DA1EC5.3020804 at usa.net> 
<46a038f90809241525t22ebb0efi6f75334516458252 at mail.gmail.com>

Today the behavior has changed a little.

First, the network problem.

The 'other host' problem turns out to be a newbie error on my part. The
router was configured as 172.18.0.1 but labeled as 172.18.0.3. We have
reconfigured the router to be 172.18.0.3 and the XO has access to the
internet via the schoolserver!

The ejabberd problem persists; however, I think it may also be a newbie
problem. I'll try to keep you posted.

Yours,

Tony

Martin Langhoff wrote:
> On Wed, Sep 24, 2008 at 11:04 PM, Tony Anderson <tony_anderson at usa.net> wrote:
>> started ejabberd:
>>
>>  chkconfig --level 345 ejabberd on
>>  service ejabberd start
>>
>> and attempted to register admin:
>>
>>  ejabberdctl ejabberd register admin schoolserver.schoolnet.gov.np admin
>>
>> This failed with a message:
>>
>> RPC failed on the node admin at schoolserver.schoolnet.gov.np: nodedown
> 
> That's strange. The procedure is right, perhaps something went wrong
> with the domain_config step? Does the /etc/ejabberd/ejabberd.cfg file
> show the correct domain?
> 
>> The command service ejabberd status gets the same error (on timeout);
>> however, ps -aux shows an ejabberd daemon.
>>
>> I then tried to set up the networks manually:
>>
>> ifcfg-eth0 was modified to use a static address (192.168.5.44). This seems
>> to work ok.
> 
> Good.
> 
>> ifcfg-eth1 was modified by adding BOOTPROTO=static. This gives an error:
>> "Error, some other host already uses 172.18.0.1."
> 
> Is that interface connected to anything? Any other host that would be
> answering for that ip address? The error you mention is from doing an
> 'arping'  to the address -
> 
> cheers,
> 
> 
> 
> m