For the TrueSight Capacity Optimization (TSCO) Gateway Server, data transfer is failing for most or all machines. The UDR Collection Manager (UCM) Status Report indicates the following transfer error repeated for each transfer attempt:

Also, the following message may be reported in the $BEST1_HOME/local/manager/log/[console]-[mgr_run]-MM-DD-YYYY-##########.log file at the time of the transfer attempt:

• TrueSight Capacity Optimization 20.02, 11.5, 11.3.01, 11.0, 10.5, 10.3, 10.0
• BMC Performance Assurance 9.5, 9.0, 7.5.10, 7.5.00, 7.4.10, 7.4.00, 7.3, 7.2

• Manager data transfer

The "Service daemon protocol terminated in the middle of session (network or disk timeout)" message indicates UDR Collection Manager on the TSCO Gateway Server and the Service Daemon on the remote agent side were able to complete their communication handshake but something then caused the data collection or data transfer request to terminate in the middle of the session.  This is usually a problem encountered on either the TSCO Gateway Server side or the TSCO Agent agent side that is considered fatal by the process which terminates the communication in the middle of the session.

To debug this problem check:

1. The Service Daemon log on the agent side ($BEST1_HOME/bgs/log/[hostname]-bgsSD.log or %BEST1_COLLECT_HOME%\bgs\log\[hostname]-bgsSD.log)
2. The udrCollectMgr log on the TSCO Gateway Server side (BEST1_HOME/local/manager/log/[console]-[mgr_run]-MM-DD-YYYY-##########.log)

User SYSTEM not authorized to pull from xxx.xxxx.xxx.xxx

If using the Authorization.cfg file to control access to the agent this error may occur if the user making the request to the Service Daemon isn't authorized.  The authorization of the user isn't done until after the communication handshake is complete and the collector or transfer request is made.  If at this point the Service Daemon identifies that the user making the request isn't authorized it will terminate the connection to the console and report the following in the log:

User SYSTEM not authorized to pull from xxx.xxxx.xxx.xxx

The TSCO Gateway Server isn't told that there was an authentication failure -- it just knows that the connection was dropped in the middle of communication which results in this error.

Warning : Unable to create directory /usr/adm/best1_default/collect/[hostname]

For example, the following message may be reported in the $BEST1_HOME/local/manager/log/[console]-[mgr_run]-MM-DD-YYYY-##########.log file at the time of the transfer attempt:
  UDR Collection Manager (11666:2) Warning : Unable to create directory /usr/adm/best1_default/collect/[hostname]

The source of this data transfer problem is that the udrCollectMgr process (which is what handles data transfer on the Linux console) is unable to create the data repository directory on your console.

The key message is from the UCM log file which shows that each transfer for each run is failing with the following
error:

This could mean a few different things:

1. The owner of the $BEST1_HOME/bgs/bin/udrCollectMgr binary is not the TSCO Agent Installation Owner user (who owns the $BEST1_HOME/collect directory)
2. The TSCO Gateway Server console UDR data repository file system is full
3. Something else is going wrong that is preventing udrCollectMgr from creating directory in your console data repository.

Check the following:

• The output of 'ls -l $BEST1_HOME/bgs/bin/udrCollectMgr' and verify that the file is owned by the TSCO Gateway Server Installation Owner and setuid (r-sr-sr-x).
• As the TSCO Gateway Server installation owner user, cd to $BEST1_HOME/collect and run 'touch test' to see if you can create a file called 'test' as the TSCO Gateway Server Installation Owner user in that directory.
• cd to $BEST1_HOME/collect and run 'df -k .' to verify that the file system isn't full.

1. BMC Performance Assurance for Servers
2. BMC Performance Analyzer for Servers