[ejabberd] removing nodes from cluster

Stephan Maihöfer sma at turtle-entertainment.de
Fri Sep 26 19:42:10 MSD 2008

right now the problem is that the database of the node was deleted
and i want to integrate it to the cluster again, using the same name.
when i try to do this from scratch i get this:
erl -name ejabberd at ejabberd2.example.com -mnesia extra_db_nodes "['ejabberd at ejabberd1.example.com']" -s mnesia
Erlang (BEAM) emulator version 5.5.2 [source] [64-bit] [async-threads:0] [kernel-poll:false]
Eshell V5.5.2  (abort with ^G)
(ejabberd at ejabberd2.example.com)1> Mnesia('ejabberd at ejabberd2.example.com'): Data may be missing, Corrupt logfile deleted: "/root/Mnesia.ejabberd at ejabberd2.example.com/LATEST.LOG", {file_error,
                                                                                                                                            "/root/Mnesia.ejabberd at ejabberd2.example.com/LATEST.LOG",
=ERROR REPORT==== 26-Sep-2008::17:38:05 ===
Mnesia('ejabberd at ejabberd2.example.com): ** ERROR ** (core dumped to file: "/root/MnesiaCore.ejabberd at ejabberd2.example.com_1222_443485_377467")
 ** FATAL ** Cannot open log file "/root/Mnesia.ejabberd at ejabberd2.example.com/LATEST.LOG": {file_error,
                                                                                           "/root/Mnesia.ejabberd at ejabberd2.example.com/LATEST.LOG",
BREAK: (a)bort (c)ontinue (p)roc info (i)nfo (l)oaded
       (v)ersion (k)ill (D)b-tables (d)istribution

i guess that stems from ejabberd1 still having some kind of "messed up" state of the database at the "old"
ejabberd2...i don't know how to fix that.
second thing is, with "cloud computing" we would bring up and down nodes as we need them and
would always have a large list of inactive nodes. apart from being kind of ugly...doesnt that somehow
slow mnesia down because it has to keep track of all the states on these nodes that are never
to be used again?


	Von: ejabberd-bounces at jabber.ru [mailto:ejabberd-bounces at jabber.ru] Im Auftrag von Zbyszek Zólkiewski
	Gesendet: Freitag, 26. September 2008 17:35
	An: ejabberd at jabber.ru
	Betreff: Re: [ejabberd] removing nodes from cluster
	this program and method was written by Alexey, so probably there isn't ;)
	anyway - what is the problem with heaving nodes inactive ?
	2008/9/26 Stephan Maihöfer <sma at turtle-entertainment.de>

		is this really the only way? shutting down ALL nodes would be very bad since we
		were thinking about turning off and on nodes as we need them (using some kind
		of cloud-computing). since then every node would be named differently we would
		end up with a list of a few hundred nodes after some time.
		there has to be a way to remove a node from the cluster in some "clean" way, isnt there?

	Zbyszek Żółkiewski

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.jabber.ru/pipermail/ejabberd/attachments/20080926/372b58bb/attachment.htm>

More information about the ejabberd mailing list