0

My hosting provider WebFusion updated my VPS with security patches late last night.

Since then I cannot start my ejabberd instance. I get the following errors, which seem to indicate that the hostname has changed.

=ERROR REPORT==== 2010-11-29 23:01:10 ===
E(<0.35.0>:ejabberd_config:278) : Error reading Mnesia database spool files:
The Mnesia database couldn't read the spool file for the table 'acl'.
ejabberd needs read and write access in the directory:
   /var/lib/ejabberd
Maybe the problem is a change in the computer hostname,
or a change in the Erlang node name, which is currently:
   'ejabberd@lvps109-104-93-116'
Check the ejabberd guide for details about changing the
computer hostname or Erlang node name.


=INFO REPORT==== 2010-11-29 23:01:15 ===
application: ejabberd
exited: {bad_return,{{ejabberd_app,start,[normal,[]]},
             {'EXIT',"Error reading Mnesia database"}}}
type: temporary

Everything has been working a treat for many months and I have made no changes to the VPS configuration, which is based on Ubuntu 8.04 LTS by the way.

I have tried everything I could find on Google w.r.t recovering or migrating mnesia files in this type of scenario, but I cannot seem to get the instance started again. I cannot make a binary backup of the files and clean-slate the system since this relies on getting a running instance up before the backup can be completed. It looks like I am stumped and will have to write-off months of user account registrations.

My hostname does not appear to have changed with the patches and reboot that the hosting company performed. Honestly, I don't really know what hostname ejabberd is expecting but I know it is of the form ejabberd@. Is there a way I can determine this from the existing mnesia data files I wonder?

As for the technical support channels of WebFusion? They have told me in no uncertain terms that I am on-my-own and offered no support at all when questioned.

Would appreciate any guidance on how to extract the data from the mnesia data files in /var/lib/ejabberd and how to recover this or a new installation. I will consider moving to a new host or VPS if I have to.

Very many thanks in advance.

2 Answers2

2
  1. Stop ejabberd.

/etc/init.d/ejabberd stop

2.Kill processes: beam, beam.smp, epmd

killall ejabberd

killall beam

killall beam.smp

killall epmd

3.Remove the files in /var/lib/ejabberd (Create a backup copy of this folder just in case)

rm -rf /var/lib/ejabberd/*

rm -rf /var/lib/ejabberd/.erlang.cookie

4.Try to start ejabberd.

/etc/init.d/ejabberd start

Ronnie
  • 21
  • 3
1

I ended up using ejabberdctl to get information about the underlying mnesia database.

Then set ERLANG_NODE within the file /etc/defaults/ejabberd to, in my case, ejabberd@vps. Et viola the service came up upon typing /etc/init.d/ejabberd start.

Upon startup the process reports failed after a period of 30 seconds. Yet I could access the web management interface and dump the database.

For future use, I shall be deployed to a new ejabberd+MySQL cluster and obviously leaving my current VPS provider.