Troubleshooting

This chapter covers how to get debugging information and troubleshoot issues in OpenAM deployments.

Solutions to Common Issues

This section offers solutions to common problems when working with OpenAM.

OpenAM Installation

OpenAM configuration could not write to the configuration directory. Where must I change permissions, and what permissions are required?
If the user running the web container has a $HOME directory, then the configuration directory is stored there, and you probably do not have this problem. If you do not know the user running the web container, use the ps command to check. In the following example, the user is mark, the web container tomcat:
$ ps -ef | grep tomcat mark 1739 1 0 14:47...
For a container installed from native packages with a dedicated user, $HOME may not be where you think it is. Look at the user’s entry in /etc/passwd to locate the home directory. The user running the web container where you install OpenAM must be able to read from and write in this directory.

If you cannot change the permissions to the user’s home directory, you can, as a workaround, unpack OpenAM-15.2.1.war, set the configuration.dir property in the WEB-INF/classes/bootstrap.properties to a directory with appropriate permissions, and repack openam.war with the adjusted file before deploying that:
$ cd ~/Downloads/openam/OpenAM-15.2.1.war $ mkdir unpacked ; cd unpacked $ jar xf ../OpenAM-15.2.1.war $ vi WEB-INF/classes/bootstrap.properties $ grep ^config WEB-INF/classes/bootstrap.properties configuration.dir=/my/readwrite/config/dir $ jar cf ../openam.war *
Deployment failed due to lack of memory. What do I do?
OpenAM requires at least a maximum heap size of 1024 MB, with a 256 MB maximum permanent generation heap size. For the Sun JVM, ensure the container starts with -Xmx1024m -XX:MaxPermSize=256m for these settings.

If you do not know the settings used when the web container was started, use the ps command to check. In the following example, the web container is tomcat:
$ ps -ef | grep tomcat | grep Xm ... -Xmx1024m -XX:MaxPermSize=256m ...
Make sure you have at least 2 GB of RAM on the system where you run OpenAM to avoid running out of memory.

If you make it through deployment and seem to be running out of memory later, you can confirm memory errors in OpenAM by searching the config-dir/openam/debug/* files for java.lang.OutOfMemoryError.

Deployment failed due to invalid hostname configuration. What do I do?

OpenAM requires that you use a fully qualified domain name (FQDN) that the host can resolve:

$ ping openam-ter.example.com
PING openam-ter (192.168.56.2) 56(84) bytes of data.
64 bytes from openam (192.168.56.2): icmp_seq=1 ttl=64 time=0.025 ms
64 bytes from openam (192.168.56.2): icmp_seq=2 ttl=64 time=0.032 ms
64 bytes from openam (192.168.56.2): icmp_seq=3 ttl=64 time=0.030 ms

For a test deployment (at home, on a laptop), you can use fake FQDNs in /etc/hosts (%SystemRoot%\system32\drivers\etc\hosts on Windows), depending on how your network is configured:

$ cat /etc/hosts | grep openam
192.168.56.2 openam openam.example.com
192.168.56.3 openam-bis openam-bis.example.com
192.168.56.5 openam-ter openam-ter.example.com

I configured OpenAM, and now am seeing the configuration screen again. Who deleted my configuration?
OpenAM uses a file in $HOME/.openamcfg/ to bootstrap and find its configuration. The file is named after the path to OpenAM and contains the path to the configuration. The following example shows what the file looks like for OpenAM deployed in Apache Tomcat under /path/to/tomcat/webapps/openam, and running as user amuser with $HOME /home/amuser:
$ cat ~/.openamcfg/AMConfig_path_to_tomcat_webapps_openam_ /home/amuser/openam
If OpenAM cannot find its configuration, then it displays the configuration screen.

OpenAM Upgrades

I have upgraded OpenAM, now my tools are not working properly. What happened?

Every OpenAM component must be upgraded, not just the main OpenAM .war file. If you did not upgrade the tools too, they may not work as intended.

OpenAM Administration

I cannot use the browser-based equivalent of ssoadm, http://openam.example.com:8080/openam/ssoadm.jsp. Why not?

For security reasons, the ssoadm.jsp page is not activated by default. For more information about activating the ssoadm.jsp page, see "OpenAM ssoadm.jsp".
The ssoadm command is very slow on my virtual machine (VMWare, VirtualBox, and so forth). How can I speed it up?

Virtual machine random devices do not always produce enough random data. The ssoadm command can hang while reading random data from the virtual machine’s random device, with the result that you can wait a minute or more for a single command to finish.

To work around this limitation on virtual machines, make sure you install something that generates enough random data, such as a timer entropy daemon.
I have OpenAM deployed on WebLogic 12.1.1 and am running Java 6. What can I do to fix the exceptions and strange results that I am seeing when I use the ssoadm command?

Edit the start up script for WebLogic as described in "Preparing Oracle WebLogic" in the Installation Guide, and then restart WebLogic.
I added OpenDJ as a data store, and now I cannot add a user. OpenAM gives me the following error
When you set up a New Data Store to use OpenDJ as an identity repository under Realms > Realm Name > Data Stores > New…, you need to check the Load schema when saved box if you want OpenAM to add the schema to OpenDJ. The box is not selected by default.

The full version of OpenAM includes directory server schema in the ~/Downloads/openam/ldif/ directory. To add the schema to OpenDJ afterwards, you can try the following command:

$ /path/to/opendj/bin/ldapmodify \ --port 1389 \ --bindDN "cn=Directory Manager" \ --bindPassword password \ --filename ~/Downloads/openam/ldif/fam_sds_schema.ldif Processing MODIFY request for CN=schema MODIFY operation successful for DN CN=schema
I have OpenAM installed in WebSphere application server with IBM Java. I am doing REST-based user registration or forgotten password reset, or setting up the HOTP authentication module, sending mail to an SMTP server over SSL.
If you see in the OpenAM Authentication debug log that the SSL handshake is failing when connecting to the mail server, then it is likely that the SSL certificate presented by the mail server is not trusted.

This is a WebSphere/IBM Java issue, rather than an OpenAM issue.

To work around the problem, follow these steps to make sure that WebSphere trusts the mail server SSL certificate:
1. Log in as administrator to WebSphere console.
2. Browse to Security > SSL certificate and key management > Manage endpoint security configurations, and then click the link for the node where OpenAM runs.
3. In the menu on the right, click SSL configuration.
4. Click NodeDefaultSSLSettings.
5. In the menu on the right, click Key stores and certificates.
6. Click NodeDefaultTrustStore.
7. In the menu on the right, click Signer certificates.
8. Click Retrieve from port.
9. Set Host, Port, and Alias, and then click Retrieve signer information.
  
  The Host is the host name of the SMTP server.
  
  The Port is the port number of the SMTP server, such as 465.
  
  The certificate Alias can be set to the user name used to authenticate to the mail server.
  
  For example, if you are sending mail through Google mail as my.user, then set Host to smtp.gmail.com, set Port to 465, and set Alias to my.user.
10. After the information is retrieved, click Apply, save your work, and then restart WebSphere.
After WebSphere restarts, it should trust the mail server SSL certificate. OpenAM therefore should be able to connect to the mail server over SSL.

For more information, see the WebSphere documentation.
My container log file is filling up with messages from OpenAM’s OAuth authorization service and OpenID Connect provider.
This behavior is governed by the log settings for RESTlet, which is used by OpenAM for OAuth 2.0 and OpenID Connect 1.0.

Use log configuration settings to turn off logging from RESTlet.

For example, if your container is Apache Tomcat (Tomcat), follow these steps:
1. Stop Tomcat:
  
  $ /path/to/tomcat/bin/shutdown.sh
2. Edit the Tomcat settings script, catalina.sh or catalina.bat, to use the logging configuration file.
  
  For example, in /path/to/tomcat/bin/catalina.sh, uncomment the following line, and then save your work:
  
  LOGGING_CONFIG= \ "-Djava.util.logging.config.file=$CATALINA_BASE/conf/logging.properties"
3. In $CATALINA_BASE/conf/logging.properties, add the following line, and then save your work:
  
  org.restlet.level=OFF
4. Start Tomcat:
  
  $ /path/to/tomcat/bin/startup.sh
I have session failover configured for an OpenAM site. I see many connections in TIME_WAIT state, and the connections seem to be used only for communication between OpenAM servers in that site. What should I set to have fewer connections in TIME_WAIT?
When you have session failover configured for a site, OpenAM servers run health checks against other servers in the same site. By default, the health checks are run every second (1000 milliseconds) with a timeout of 1 second (1000 milliseconds).

If there is network latency between servers in a site, for example, if you are running your servers in virtual machines, the default settings might not be right for your deployment. In that case, consider changing the following advanced server properties:
- By lengthening com.iplanet.am.session.failover.cluster.stateCheck.timeout and com.iplanet.am.session.failover.cluster.stateCheck.period to something longer than the default, you can work around issues with network latency.
- By setting com.sun.identity.urlchecker.dorequest to true or false, you can change whether OpenAM performs an HTTP GET request or only checks the Socket connection of com.sun.identity.urlchecker.targeturl as a health check.
To set advanced properties, either use the OpenAM console page under Configure > Server Defaults > Advanced, or set the properties using the ssoadm update-server-cfg command as in the following example, which updates the default server configuration:
$ ./ssoadm \ update-server-cfg \ --servername default \ --adminid amadmin \ --password-file /tmp/pwd.txt \ --attributevalues com.iplanet.am.session.failover.cluster.stateCheck.timeout=2000
I want to change the password for the UrlAccessAgent that was assigned during the OpenAM Installation process.
To change the UrlAccessAgent password, you can use the ssoadm that is installed with the OpenAM Administration tools, as described in the "Installing OpenAM Tools" in the Installation Guide

You can then change the password for the UrlAccessAgent for the root realm (/) with the following commands:
$ cd /path/to/SSOAdminTools/bin $ ./ssoadm \ set-identity-attrs \ --realm / \ --idname amService-UrlAccessAgent \ --idtype user \ --adminid amadmin \ --password-file /tmp/passwd \ --attributevalues userpassword=changeit Attribute values of identity, amService-UrlAccessAgent of type, user in realm, / was modified.
The new password will take effect the next time you start OpenAM.
How do I set the TCP connection timeout for outbound HTTP connections created by OpenAM?

The default connection timeout is 10 seconds (10000ms). You can change the default timeout by setting the org.forgerock.openam.url.connectTimeout property.

On the OpenAM console, go to Configuration > Server & Sites > Server Instance > Advanced, and then set the property with a value representing the TCP connection timeout in milliseconds.