Event storage is also known as the ring buffer. You can set the event storage use limit and the storage mode for the resource use policy for all of your host endpoints or selected host sets using the Web UI or the API.
The event storage limit represents the number of megabytes of storage that can be used for tasks performed on behalf of the Endpoint Security (HX) server.
If the agents' events database exceeds the storage specified by this setting, the agent rebuilds the database. If the database is corrupt and the attempt to rebuild it fails, a new, empty events database is created. The old events database is backed up as file events.db.bk. You may need assistance from Trellix Technical Support to extract the data contained in the backup.
The event storage mode allows you to adjust event storage I/O handling and physical disk I/O use. If your host endpoints experience a degradation in I/O processing times, use this setting to optimize event storage I/O handling and reduce the physical disk I/O use. The default settings is conventional.
Important
Changing storage mode may lead to the loss of historical data. Complete all incident response investigations on your host endpoint before changing the storage mode.
The storage_mode setting is only available for agents running Trellix Endpoint Security (HX) xAgentversion 21.33.7 or later.
In Endpoint Security (HX) xAgent version 26, databases operate more reliably and allow for automatic recovery whenever possible.
The table below describes the available storage modes.
Storage Mode | Description | Configuration |
|---|---|---|
Conventional | This is the default storage mode. This storage mode stores write events in batches before writing them to the events database on disk. ImportantUse one of the other storage modes only if your endpoints experience a degradation in I/O processing times. | To select the default storage mode, remove the |
Memory-mapped I/O with log | This is the recommended memory-mapped storage mode. In this storage mode, read and write operations are backed by memory and backed by an on-disk journal file. Write transactions are written to the Write Ahead Log file ( Memory-mapped I/O with log mode improves disk I/O performance for two reasons.
Before changing to this storage mode in your production environment, you should first determine the impact of the change in a test environment. | To select this storage mode, set the storage_mode key to |
In-memory | This storage mode stores the events database in memory. This mode increases memory usage, but improves disk I/O performance. Consider the following caveats before switching to in-memory storage mode.
| To select this storage mode, set the |
Memory-mapped I/O NOT RECOMMENDED | Use of this storage mode is not recommended because it may cause failure of Real-time Event storage and functionality. Please use MMAPIOWITHLOG. This storage mode uses memory mappings to back up read and write operations, and uses paging to optimize the read and write operations. | To select this storage mode, set the |
This section covers how to set the event storage use limit using the Web UI. See the Endpoint Security (HX) REST API Guide for more information on setting the event storage use limit using the API.