Search <book_title>...

Enterprise Vault™ Introduction and Planning

Last Published: 2021-12-06

Product(s): Enterprise Vault (14.2)

How to plan installing the Indexing Service

There must be at least one Indexing Service in an Enterprise Vault site. However, there can be only one Indexing Service on a computer. If you have more than one Indexing Service in an Enterprise Vault site they must run on separate computers. It is usually best to install the Indexing Service on the same computer as the Storage Service.

Note:

For the Indexing Service to function correctly, you must install it on a computer whose year format is equivalent to the Gregorian year (currently 2021). For example, setting the computer's regional format to Thai causes the Indexing Service to fail, as the current year in the Thai solar calendar falls outside the date range that the service supports.

In larger or distributed deployments, consider using Index Server groups to spread the indexing load. Ensure that associated Storage and Indexing Services are either collocated, or can communicate over fast connections.

An Indexing Service can manage simultaneously the indexes for many archives, which may be stored in different vault stores on different Storage Service computers. The index locations assigned to the Indexing Service must have sufficient disk space to store the indexing data.

The indexes are organized as follows:

There is a separate Elasticsearch index for each archive.
Each Elasticsearch index can contain data from multiple archives.
Elasticsearch index can get rolled over when the following thresholds meet:
- Archive count per index
- Size limit of total shards
Taking snapshots is the recommended backup method for Elasticsearch indexes.
For more information, see https://www.veritas.com/support/en_US/article.100051446.

You specify during configuration the index location where index volumes are to be created. From Enterprise Vault 14.2 and later, you can specify only one index location. It is recommended that you configure a index location which is expandable. (This location is sometimes referred to as "index root path" in error and diagnostic messages.)

If you create Index Server groups, then indexing the associated vault stores is shared by all the Index Servers in the group. The index volumes are spread over the locations assigned to the Index Servers in the group.

You can choose how much information is indexed for items in an archive using the indexing level; this can be brief, or full indexing. If you want to be able to search item content for phrases, for example using Compliance Accelerator or Discovery Accelerator, then you need to use full indexing.

The more information that is indexed about an item, the easier it is to search for it. However, the more information that is indexed about an item, the more disk space is required for the index. The size of index data for an item varies with the indexing level.

Table: Estimated size of indexing data shows the estimated size of an index as a percentage of the unarchived size of the item for the different indexing levels.

Table: Estimated size of indexing data

Indexing level	Estimated size
Brief	4%
Full	8%

So, if you have allowed in the region of hundreds of gigabytes (or terabytes) of space for your vault stores, you are likely to require in the region of gigabytes (or tens of gigabytes) for your indexes.