eEM users sometimes cannot login to AAI when primary eEM server is down.

Products

• Automic Automation Intelligence

Issue/Introduction

In and environment with eEM failover configured, eEM users sometimes cannot login to AAI when primary eEM server is down.

They will attempt to login and it will take a long time for the login to eventually fail.

Environment

Release : 6.2.0

Component : AUTOMIC AUTOMATION INTELLIGENCE ENGINE

Resolution

Upgrade to at least 6.4.1 which introduced several new parameters, which should prevent AAI from trying to connect to the primary eEM server for longer then 30 seconds by default before attempting a secondary eEM server.. 

Additional Information

The 3 new server parameters introduced int 6.4.1 are:

1. eEM.connTimeout: The time to wait for a connection to be established to an eEM server. If the connection to eEM server is slow, this value should be increased. 
2. eEM.retryCount: Number of times to retry a connection to an eEM server before failover to next eEM server. Increase the count if the network to eEM server is not stable.
3. eEM.log4jConfigFile: logging config file for eEM client. A default eem.log4j.xml is available in AAI configuration directory. The parameter is to be set if there is eEM problem so that we can collect eem logging, restarting server is required to make the file effective. It will create log files: eiam.sdk.java.log and eiam.network.java.log. To enable debugging, change levels to TRACE in the xml file. The file will be reloaded every minute once the file was in use.

To turn on eem logs "eem.network.java.log" and "eem.sdk.java.log" update eem.log4j.xml by updating the "level value" to TRACE for Network and SDK.  

 

 

 <!-- valid levels are: ALL < TRACE < DEBUG < INFO < WARN < ERROR < FATAL < OFF-->
   <logger name='Network' additivity='false'>
       <level value='TRACE'/>
       <appender-ref ref='Network' />
   </logger>
   <root>
       <level value='TRACE'/>
       <appender-ref ref='SDK' />
   </root>
</log4j:configuration>

 

Existing server parameters usage:

1. eEM.connectionTimeout: Maximum time AAI will wait for a connection to be established to an eEM server or user authentication to complete, before the login request that initiated the connection is failed. Note for eEM failover to work, the value has to be at least eEM.connTimeout * (1 + eEM.retryCount) + eEM.socketTimeout."  This value should be higher if the expected number of concurrent web UI users is high.

2. eEM.socketTimeout: the time to wait for a packet to be received.