Jump to: navigation, search


LFMT Database Sizing

The LFMT Database is largely sized based on two tables.

  • The log_file table holds the names and locations of the log files stored in CLS.
  • The log_file_indexed_token table holds any custom index reference defined by the customer for specific application types.

LFMT Database - Detailed Sizing Approach

Database sizing can be calculated using the following metrics:

  • 0.50kB per log file
  • 0.15kB per index

Sample LFMT Database Sizing

Assume:

  • CLS Total Disk Consumption = 1TB
  • CLS Total Files = 1000000
  • Indexes per file = 50
Important
The above values are the equivalent to 10MB log files compressed at a 10-1 ratio.

Sample calculation:

log_file table size = 1000000*.50kB = 500MB

log_file_indexed_token table size = 1000000 * 50 * 0.15kB = 7324MB

Total DB size = 500MB + 7324MB = 7824MB = 7.7GB

LFMT Database - Simplified Sizing Approach

Use the following table to estimate the DB space requirements for the LFMT Database. Average daily call volume can be used as a reference. Note that a deployment would typically have a separate LFMT Server host per datacenter so the sizing should represent the call volume handled by that data center.

Average Daily Calls Database Space
Less than 10k 500MB
Between 10k-30k 1GB
Between 30k-120k 30GB
Greater than 120k Please consult with Genesys Customer Care. (Detailed Sizing approach recommended)
Important
  • The DB space utilized by LFMT is highly dependent on the number of indexes configured per file type. When using the Simplified LFMT sizing method, we recommend you start by configuring a low log file retention period (e.g.: 3 or 4 days) and then adjust it accordingly after data collection is started and it is clear there is sufficient DB space to handle all the indexes, or adjust the space accordingly.
  • The maximum value for the ID is 2,147,483,647 for both the LOG_FILE and LOG_FILE_INDEXED_TOKEN TABLES. Please periodically monitor the value of the ID in both the tables and if it is approaching the said maximum value, the ID will need to be manually re-seeded (via the sequence for Oracle and PostGres, and via the Identity Column for MSSQL.

Feedback

Comment on this article:

blog comments powered by Disqus
This page was last modified on July 17, 2018, at 09:12.