Why do I get the error “File-set is exhausted, recording will stop (no more space available in fileset and rollover inactive)”?

The error message “File-set is exhausted, recording will stop (no more space available in fileset and rollover inactive)” comes from the fact that the default maximum size for a file segment is 2 GB. When the recorded segment reaches 2 GB, Recorder stops.

Below are two options for avoiding this limitation, along with their advantages and disadvantages. You will need to analyze which is more suitable for each use case, taking into account your system's resources. For more detailed information, see Section 4.5 “Database (Output File) Properties” in the “RTI Recording Service User’s Manual” for RTI Connext DDS 5.2.3.

Option 1) Increase max_file_size from 2 GB to a greater value.

<recorder_database>
    <max_file_segments>DDS_Long<max_file_segments>
<recorder_database>
  • Advantages: You will have the recorded samples saved in the same database file segment.
  • Disadvantages: The size of this file segment could be huge. Since you are publishing large samples at a high rate, you may reach the new upper limit and once again exhaust the file segment size. 

Option 2) Increase max_file_segments to force Recorder to create a new segment when the current one reaches the 2 GB of max_file_size.

  • Advantages: You can set a very high value and be sure that when a segment reaches the max_file_size limit, a new segment will be created.
  • Disadvantages: You will have several large file segments. To make space on the disk for new segments, you may need to save the oldest ones as a backup. However, once Recorder reaches the max_file_segment value, it will stop again. You can avoid this by setting rollover to true. Then Recorder will start again using segment 0 when it reaches the max_file_segments already recorded. Note: The samples recorded in the “new” segment 0 will have a newer timestamp than the existing ones in the following segments. This can cause Replay to fail, since it will expect to find a newer timestamp in following segments, not an older one.