Open topic with navigation
The DIH distributes index actions to its child servers.
In standard mirror and non-mirror mode, DIH forwards the IDX or XML index file to all its child servers, with instructions to the child servers about which documents it must index. You can use different modes in the DIH to reduce the amount of data that the DIH, or the child servers, must process.
The DIH configuration parameter
PreserveDREADD allows it to distribute index actions more quickly.
By default, DIH accepts an IDX or XML file input and converts the original
DREADD index action into a
DREADDDATA index action, which contains all the IDX or XML data. If you set
true in the configuration file, DIH forwards the
DREADD action directly. This option reduces the network load.
To use this parameter, each child IDOL Content component must be able to access the original IDX or XML file.
You can use
PreserveDREADD only in mirror mode, simple non-mirror mode or
DistributeOnBatch mode. For example, you cannot use
PreserveDREADD if you have set
true. In advanced distribution modes, DIH must parse the IDX or XML file to distribute the documents correctly.
In distribute on batch mode, DIH indexes data to alternating child servers. This method is an efficient way of distributing the data to child servers in non-mirror mode. You can use this method if:
you index a steady stream of IDX or XML files, with similar numbers of documents. This mode might not be appropriate if you infrequently index very large IDX or XML files.
you do not need to remove duplicate documents.
DIH has several distribution modes that determine how it distributes documents between child servers in non-mirror mode. Unlike standard non-mirror mode, these methods send each child server only the documents that it must index. This option reduces the amount of data that each child server receives.
If you use these methods, the number of DIH child servers is fixed. To add child servers, you must clear all engines, add the new servers, and then re-index using the DIH.
In distribute by fields mode, DIH sends documents with the same value of the specified field to the same child server. DIH determines which child server receives a given field value.
In distribute by field values mode, you can specify which child servers receive documents that contain certain field values.
The Round-Robin indexing mode rotates indexing between several child servers, so that indexing and querying are performed by different child servers at different times.
The IDOL Content component provides the fastest queries when it is not indexing, and indexes fastest when it is not being queried. When you configure indexing for round-robin mode, DIH suspends query handling for a specific child server. It then indexes only into this child server, which has optimal indexing performance.
After the specified time period, DIH starts indexing to a different child server and makes the previous server available for querying.