powered by Jive Software

Invalid username or password after upgraded to Openfire 3.4.4

Users are no longer able to login into Spark since I upgraded to Openfire 3.4.4. Has anyone had this problem?

What did you upgrade from? Also, this sounds odd, but are you using a compound user or group filter (ie, there is an ‘&’ in your filter)?

We upgraded from version 3.3.3. We are AD integrated.

That is what I thought… There is a bug that was introduced in 3.4.x with the XML parser in the setup program of Openfire. Because the XML is misformed, it doesn’t pull users properly once you restart the service.

If you use the ‘&’ charater in your group or user filter statement, the XML parser in the setup program doesn’t parse it correctly… the xml will double-up the & escape sequence, so you will need to manually edit these out.

For example, our filter for users is:


This should show up under c:\progfiles\openfire\conf\openfire.xml as:

<searchFilter>(&amp;(objectClass=user)(extensionAttribute10=jabber))&l t;/searchFilter>

but the XML parser in the configuration program screws up and puts it as:

<searchFilter>(&amp;&amp;(objectClass=user)(extensionAttribute10=j abber))</searchFilter>

To fix, stop the service, edit your user and group filters in openfire.xml, save your changes, and restart the service. I hope that does the trick for you.

I have noticed the same thing:


Actually, it appears that the XML I posted in from my config file was

messed up when the forum escaped the text… Oh well. Have you heard

from any developers or anyone at Jive about this issue? Do you know if

there is an issue created for this? I don’t appear to have permissions

to create issues or I would just create one.

According to Jadestorm it has been fixed with JM-703

I changed the search filters as you listed and users were able to login. After users logged out and back in “invalid username and password” comes up. Any suggestions on how to fix this? Should I have to make any changes on Opefire Admin Console?

I upgraded to Openfire 3.4.5 and Spark 2.5.8 and it seems to be working ok now