Bug 817997 - osa-dispatcher service will not start unless "/var/lib/jabberd/db/" is cleared out beforehand
Summary: osa-dispatcher service will not start unless "/var/lib/jabberd/db/" is cleare...
Alias: None
Product: Spacewalk
Classification: Community
Component: Server
Version: 1.7
Hardware: x86_64
OS: Linux
Target Milestone: ---
Assignee: Milan Zázrivec
QA Contact: Red Hat Satellite QA List
Depends On:
Blocks: space18
TreeView+ depends on / blocked
Reported: 2012-05-02 01:29 UTC by Dan Burkland
Modified: 2012-11-01 16:23 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Last Closed: 2012-05-08 12:18:13 UTC

Attachments (Terms of Use)

Description Dan Burkland 2012-05-02 01:29:41 UTC
Description of problem: The osa-dispatcher service will not start unless "/var/lib/jabberd/db/" is cleared out beforehand

Version-Release number of selected component (if applicable): osa-dispatcher-5.10.41-1.el6.noarch (Spacewalk 1.7)

How reproducible: Everytime

Steps to Reproduce: /etc/init.d/osa-dispatcher restart (or start)
1. /usr/sbin/spacewalk-service restart (or start)
Actual results: The init script shows that the osa-dispatcher service starts however a "/etc/init.d/osa-dispatcher status" shows that it is "dead". A simple "ps" shows that indeed the service is no longer running

Expected results: It should start without issue

Additional info: I tailed the "/var/log/rhn/osa-dispatcher.log" file and noticed the following entries after running the "/etc/init.d/osa-dispatcher start" command:

2012/05/01 20:19:58 -05:00 31439 osad/jabber_lib.__init__
2012/05/01 20:19:58 -05:00 31439 osad/jabber_lib.setup_connection('Connected to jabber server', 'srv03.dburkland.com')
2012/05/01 20:19:58 -05:00 31439 osad/jabber_lib.register('ERROR', 'Invalid password')

If I clear out the "/var/lib/jabberd/db/" directory and then start the osa-dispatcher service it works fine.

Comment 1 Jan Pazdziora 2012-05-03 11:45:30 UTC
(In reply to comment #0)
> If I clear out the "/var/lib/jabberd/db/" directory and then start the
> osa-dispatcher service it works fine.

So by clearing out /var/lib/jabberd/db, were you able to resolve the issue on your installation for good, or does the problem reoccur?

How many clients using osad do you have? Is the auth database in good shape (not corrupted, etc.)?

Comment 2 Dan Burkland 2012-05-04 05:39:57 UTC
Before each restart I must clear /var/lib/jabberd/db otherwise the osa-dispatcher service will fail to start (it reoccurs).

I have about 6 clients via OSAD (just a small home network). As far as I know the auth database is in good shape and not corrupted but wouldn't know how to find out for sure.

Thanks for the reply,


Comment 3 Milan Zázrivec 2012-05-04 07:10:46 UTC
OK, please connect to your database via sqlplus (if Oracle) or psql
(if PostgreSQL) and see what the following query returns:

    select id, jabber_id, password, hostname from rhnPushDispatcher;

Then, please paste the output from

    tail -f /var/log/messages

while doing

    service jabberd restart

Comment 4 Dan Burkland 2012-05-05 07:59:23 UTC
The posted query returns the following results: http://pastebin.com/kxCcdRZy

"/var/log/messages" output from when I ran the "service jabberd restart" command: http://pastebin.com/Yzz6iqgw

It looks like I have an incorrect hostname somehow in my Spacewalk database, any idea on how to delete that record?



Comment 5 Milan Zázrivec 2012-05-06 18:18:49 UTC
Your /var/log/messages look all right, how you got to have two entries
in rhnPushDispatcher -- I don't understand.

Nonetheless, try to connect to your database with psql again and delete
the incorrect fqdn with:

    delete from rhnPushDispatcher where id = 1;

restart your Spacewalk instance with:

    spacewalk-service restart

and see whether this helps your situation.

Comment 6 Dan Burkland 2012-05-06 23:55:56 UTC
I ran the suggested commands and while the osa-dispatcher service stayed running, several error log entries would appear each time I restarted the osad service on my Spacewalk clients. I resolved this by:

1) Again clearing out /var/lib/jabberd/db
2) Restarting the Spacewalk services on the server several times to make sure it wasn't a fluke
3) Restarting the osad service on my Spacewalk clients several times

After performing the aforementioned steps no error messages appear in the logs and everything seems to work just fine. 

One thing I forgot to mention was that the hostname of this system recently changed and I'm guessing the removal of the previously mentioned database record was skipped over by the spacewalk-hostname-rename tool.

Thanks again for your help!


Comment 7 Milan Zázrivec 2012-05-08 12:18:13 UTC
Great, since the problem described here seems to be resolved, I'm closing
this report.

Note You need to log in before you can comment on or make changes to this bug.