Hi !
I am using fedora 13 ejabberd-2.1.5-1.fc13.i686 with MySQL. The connections to the mysql DB is fine from both the nodes.
I have followed step by step. But when I start the server on node 2 after doing
"mnesia:change_table_copy_type(schema, node(), disc_copies)."
I get the following error in ejabberd.log
=ERROR REPORT==== 2010-08-17 19:15:09 ===
E(<0.38.0>:ejabberd_rdbms:67) : Start of supervisor 'ejabberd_odbc_sup_jabberm.xyz.com' failed:
{error,{{'EXIT',{badarg,[{ets,delete,[sql_pool,"jabberm.xyz.com"]},
{mnesia,delete,5},
{mnesia_tm,non_transaction,5},
{ejabberd_odbc_sup,start_link,1},
{supervisor,do_start_child,2},
{supervisor,handle_start_child,2},
{supervisor,handle_call,3},
{gen_server,handle_msg,5}]}},
{child,undefined,'ejabberd_odbc_sup_jabberm.xyz.com',
{ejabberd_odbc_sup,start_link,["jabberm.xyz.com"]},
transient,infinity,supervisor,
[ejabberd_odbc_sup]}}}
Retrying...Node 1 is running fine with mysql!
If I do not run the steps >"mnesia:change_table_copy_type(schema, node(), disc_copies)."< I can run the node2 server but, it shows up as stopped nodes in the node1 admin interface.
I am able to run the node1 as stand alone[ with mysql] without any problems. But I need ejabberd clustered.
Please help as I dont know how to debug this error.
Thank you!
Anand.
Re: Clustering + mysql
I'm having the same problem. I've build my 2.1.5 from source. If I find an answer I'll post. On my 2.0 cluster, I don't see that sql_pool field when I run mnesia:info().
Re: Clustering + mysql
Something was broken. To fix, I deleted the contents of the whole $PREFIX(for me that was /opt/ejabberd-2.1.5) and did make install from source. Then I noticed that I MUST qualify the cookie on the command line(for Guide direction 6.2.2). I also qualified full names and not short names.Then everything worked fine. Not the best fix(delete everything), but I had to get this done.