Talk:Installing ejabberd: Difference between revisions

From OLPC
Jump to navigation Jump to search
 
(8 intermediate revisions by 4 users not shown)
Line 43: Line 43:


I am using Ubuntu Gutsy and have installed my ejabberd server on my laptop from your excellent instructions. The one other thing worth mentioning for folks is that the special instructions for creating the .pem file are still necessary with this approach. It isn't there until you do. Maybe these should go in here, making it one-stop shopping?
I am using Ubuntu Gutsy and have installed my ejabberd server on my laptop from your excellent instructions. The one other thing worth mentioning for folks is that the special instructions for creating the .pem file are still necessary with this approach. It isn't there until you do. Maybe these should go in here, making it one-stop shopping?
:Tested and added. --[[User:Morgs|morgs]] 12:05, 16 July 2008 (UTC)


== Ejabberdctl status fails with "RPC failed on node status@servername: nodedown" . ==
== Ejabberdctl status fails with "RPC failed on node status@servername: nodedown" . ==
ridd 18 May 2008. Edited to remove extraneous. ridd 2008-08-02
Ejabbered is running (started with ejabberd start) as logs show my attempts at connecting via
port 5280. Expect this is a problem with -sname or elang cookie but can not find source of any
conflict. Need ejabberdctl to set admin password. This is install from OLPC stable repository -
beta 2.0.0 running on CentOS 5 VPS (no errors on install). All version requirements meet for
openssl, expat, erlang. Install/configure instructions followed. Have tried hosts of domainname
and "localhost" and adding to /etc/hosts file (which contains automatically added IP servername).
Source of conflict ?? ridd 18 May 2008.


Solution: Must include node name in running ejabberdctl. In my case with servername of kitty:
Solution: Must include node name in running ejabberdctl. In my case with servername of kitty:
Line 59: Line 54:
Only after ejabberd is started (/etc/init.d/ejabberd start) can ejabberdctl be used (can not be used
Only after ejabberd is started (/etc/init.d/ejabberd start) can ejabberdctl be used (can not be used
to start). ridd 23 may 08
to start). ridd 23 may 08

:Strange... what distro are you running on?
:BTW, you can sign your comments automatically by typing <pre>--~~~~</pre> --[[User:Morgs|morgs]] 14:34, 22 May 2008 (EDT)

::From school repository, compiled binaries, testing repository, rpm: ejabberd-2.0.0-0.1.beta1.olpc.i386.rpm dated 12 Mar 2008. Also, for wiki entry for Y server, CentOS SME, cautions not to use ejabberdctl for starting/stopping. This is running on CentOS 5 on a virtual server. Note node form options at end of ejabberdctl.cfg -- all options commented out -- appears to default to second option of ejabberd@servername. . ridd 31 may 08.

See also [[Ejabberd Configuration#Troubleshooting]] - the nodedown error also occurs when changing hostname. Remove your configuration and reinstall. --[[User:Morgs|morgs]] 12:10, 16 July 2008 (UTC)

Noted on [serv-devel]"Jabber now working" Brian Berry 10, jul 2008:
:He changed /etc/hosts, servername on localhost ip. My servername was on exposed ip. Solution ?
ridd 27 aug 2008

== Removing User Results in repeated "failed authentication" on Next Visits. ==

Accidentally removed my own XO and it just would not re-register. Corrected, but specific fix unkown.

Latest revision as of 18:36, 6 December 2008

Instructions for Fedora

I'd like to have these instructions on building ejabberd for Fedora too, but I don't have Fedora. Perhaps someone can do a Fedora version?--morgs 08:52, 4 March 2008 (EST)

Packages

RPMs are currently a bit out of date (ejabberd2 beta). debs are very out of date (ejabberd 1.4). If/when I can build newer packages and/or prod those who built the current packages, I'll add them as installing packages would be much easier than building from source.--morgs 08:54, 4 March 2008 (EST)

Can't see the "everybody" group?

I followed these instructions as well as I could, but I don't see the "everybody" group in Gajim. The sugar-jhbuild's name is "sugar1". When I try "ejabberdctl connected-users", I just get the Gajim client connected.

Suggestions?

You need at least two users on the server for it to show up on your buddy list. You don't see yourself. What works for me, although it's a bit confusing, is to add a second account in the same client (I used pidgin in my testing, but should be the same in others) - then put both accounts online, and the "everybody" group appears and shows both accounts. (What is really happening is that one account sees the other, and the second account sees the first, in the same buddy list.)
Or, you could use two different jabber clients on the same machine - less confusing but more work...--morgs 03:08, 6 March 2008 (EST)
I should be more exact. Basically, I'm trying to set up a development environment for a Sugar activity. I have two Sugar profiles - "sugar1" and "sugar2". When I share a Sugar activity from "sugar1", it does not show up in either the Analyze activity OR the other sugar-jhbuild environment. I have the jabber server set to "localhost" on both profiles. Suggestions? --kawk 21:37, 7 March 2008 (EST)
Time to look at the presenceservice.log on those sugar instances - in ~/.sugar/sugar1/logs/presenceservice.log and .../sugar2/... - if you can pastebin those or stick them somewhere, I'll take a look.--morgs 06:31, 8 March 2008 (EST)
The result is exactly the same on both profiles:
1205969172.661917 INFO s-p-s: Starting presence service...
1205969172.690407 DEBUG s-p-s.psutils: IP4 address now '136.159.209.198' (eth1)
1205969172.722435 DEBUG s-p-s.presenceservice: Debug flags: set([])
1205969172.723361 DEBUG s-p-s.telepathy_plugin: <ServerPlugin object at 0x8372464 (telepathy_plugin+TelepathyPlugin at 0x8350b60)>: Starting up...
1205969172.798689 DEBUG s-p-s.telepathy_plugin: <LinkLocalPlugin object at 0x8372784 (telepathy_plugin+TelepathyPlugin at 0x8350bc0)>: Starting up...
1205969172.798884 DEBUG s-p-s.telepathy_plugin: <LinkLocalPlugin object at 0x8372784 (telepathy_plugin+TelepathyPlugin at 0x8350bc0)>: Postponing connection
1205969172.803628 DEBUG s-p-s.buddy: Successfully preloaded buddy props
1205969172.833682 DEBUG s-p-s.telepathy_plugin: <ServerPlugin object at 0x8372464 (telepathy_plugin+TelepathyPlugin at 0x8350b60)>: connecting...
1205969172.834047 DEBUG s-p-s.telepathy_plugin: <ServerPlugin object at 0x8372464 (telepathy_plugin+TelepathyPlugin at 0x8350b60)>: Connect() succeeded
Suggestions? Maybe we can set up a IRC time? kawk 19:36, 19 March 2008 (EDT)
Sure, I'm morgs on freenode in #olpc. I'm at GMT+2, will be around this evening so until 19:00 GMT.--morgs 06:19, 20 March 2008 (EDT)
I'm kawk on irc.freenode.net#olpc, I'll be online from now until 20:30 GMT. I'll be checking back every fifteen minutes or so. Thanks. kawk 12:43, 21 March 2008 (EDT)
Never mind, I got it. You need to have the conference.your.server.name entry in /etc/hosts.

PEM creation still needed with this approach

I am using Ubuntu Gutsy and have installed my ejabberd server on my laptop from your excellent instructions. The one other thing worth mentioning for folks is that the special instructions for creating the .pem file are still necessary with this approach. It isn't there until you do. Maybe these should go in here, making it one-stop shopping?

Tested and added. --morgs 12:05, 16 July 2008 (UTC)

Ejabberdctl status fails with "RPC failed on node status@servername: nodedown" .

ridd 18 May 2008. Edited to remove extraneous. ridd 2008-08-02

Solution: Must include node name in running ejabberdctl. In my case with servername of kitty: = /usr/sbin/ejabberdctl ejabberd@kitty status (example for getting status)

Only after ejabberd is started (/etc/init.d/ejabberd start) can ejabberdctl be used (can not be used to start). ridd 23 may 08

Strange... what distro are you running on?
BTW, you can sign your comments automatically by typing
--~~~~
--morgs 14:34, 22 May 2008 (EDT)
From school repository, compiled binaries, testing repository, rpm: ejabberd-2.0.0-0.1.beta1.olpc.i386.rpm dated 12 Mar 2008. Also, for wiki entry for Y server, CentOS SME, cautions not to use ejabberdctl for starting/stopping. This is running on CentOS 5 on a virtual server. Note node form options at end of ejabberdctl.cfg -- all options commented out -- appears to default to second option of ejabberd@servername. . ridd 31 may 08.

See also Ejabberd Configuration#Troubleshooting - the nodedown error also occurs when changing hostname. Remove your configuration and reinstall. --morgs 12:10, 16 July 2008 (UTC)

Noted on [serv-devel]"Jabber now working" Brian Berry 10, jul 2008:

He changed /etc/hosts, servername on localhost ip. My servername was on exposed ip. Solution ?

ridd 27 aug 2008

Removing User Results in repeated "failed authentication" on Next Visits.

Accidentally removed my own XO and it just would not re-register. Corrected, but specific fix unkown.