Running Indexing

During indexing, CloudView proxies assign each object extracted from your database to a content engine according to its reference ID. When you enter a query, the CloudView proxies query all slices and deliver a single list of results.

Once indexing is complete, check indexing status and indexed objects.

Before you begin: Check that:
  • Your config.xml contains the correct parameters for your environment. If you are updating your index with a specific type or field, add it to the config.xml file first.
  • Each content engine to use is running. If one of the slices is down, indexing proceeds as usual and search results do not include any information from that content engine.
  • The MCS (the main Live Collaboration Server) and each FCS with files to index are running.
  • The MQL client running the indexing process is on the same time zone as the database server and the Advanced Search Server.
Configure the Java options for Studio Modeling Platform on the machine running the indexing process. To do so, define:
  • MX_JAVA_OPTIONS in the enovia.ini file on Windows.
  • mxEnv.sh (or other startup script) on UNIX.
See Also
Limitations with EXALEAD File Indexing
Scheduling Partial Indexing
Clearing the Index
Disabling File Content Indexing
Clearing Checkpoint Records
Pseudonymizing User Data
  1. Using MQL, specify the context and apply the config.xml file.

    For more information, see Applying configuration.

  2. Run the command for your indexing process (for more information, see Indexing Options):

    Indexing Process MQL Command
    Baseline (to run for the first indexing) start searchindex mode FULL;
    Partial start searchindex mode PARTIAL;
    Update udpate searchindex type TYPE field FIELD [files];

  3. Check indexing status using the following MQL command: 

    status searchindex;

    You can see the status of each vault (for more information, see Display Index Statistics).
  4. Check indexed objects and relationships using the following MQL command: 

    validate searchindex [vault PATTERN] [type|relationship] [interval START_TIME END_TIME];

    For more information, see Count and Verify Missing Objects.

    Recommendation: Do not run this command with each partial indexing. Run it once a day or week when there is less activity on the platform.

Warning: When you use MQL to run an indexing process, harmless Oracle connection warnings may appear in the mxtrace.log. These warnings appear even when the MX_CONNECTION_POOL_SIZE parameter specifies a number greater than the number of threads in config.xml. This is because the indexing integration uses 5 threads in a thick client. The thick client reads MX_CONNECTION_POOL_SIZE parameter, but unlike a server, it uses this setting only when the kernel returns a connection to the pool. The thick client initializes with 1 connection (despite a setting of 10 for MX_CONNECTION_POOL_SIZE) and the indexing integration asks for 4 more, which produces 4 harmless Oracle connection warnings.