Considerations for continuous index failure

The Order Search or the index in Elasticsearch might occasionally have failures. You must monitor your search indexes as the program functions even in the event of a problem.

The search index is non-critical for business functionality. Therefore, Sterling Order Management System Software continues to function correctly even during prolonged problems with the index operations. However, this may cause the following problems:

  • For every index failure, an exception is logged through ‘Exception Notification’, too many of these problems are logged continuously.
  • If indexing does not work for a long time, the search results from Order Search becomes unreliable as continuous failure of the indexing operation has resulted in a stale index data.

For these reasons, determine if there is a continuous problem with index operations, disable the operation proactively, and alert the users.

Such a determination is made by using the ‘fail fast’ logic. You can enable or disable this logic through the yfs.ssi.fail.fast yfs property. By default, it is enabled. This logic ensures that upon continuous index operation failures, the operation is disabled until you fix the underlying problem and re-enable the operation. If there is a continuous failure with the indexing operation, the indexing operation is disabled, and no more attempts are made to index.

The definition of what constitutes a ‘continuous’ failure is configurable through a set of properties. For more information on how to configure this feature, see properties that start ‘yfs.ssi.fail.fast’.

Note:
  • The YFS_Index_Status table of ‘STATISTICS’ table type is used to track the status of index operations by using the IndexWorking flag. When the index operation needs to be disabled, the IndexWorking flag is set to ‘N’. This data is tracked separately for the order and shipment indexes.
  • For an index, when indexing is disabled (IndexWorking is set to ‘N’) due to continuous failure of indexing operation, it implies that the index is now stale and the search results from Order Search becomes unreliable.
  • You can manage the IndexWorking flag through the Index Management Console in System Management Administrator.
  • When a JVM detects continuous failure of index operations, it notifies you through alerts and events before disabling the corresponding flag.
  • When the index operation is disabled, fix the underlying problem and enable the operation again through the Index Management Console.
  • While the index operation is disabled, every attempt to index displays a warning in the application logs.

The SearchWorking attribute in the YFS_Index_Status table is not used in the built-in integration with Order Search.