End-of-Life (EoL)

Disaster Recovery and Live Backup Overview

Describes live backup, how to configure your environment, server DR status, and DR scenarios.
Live Backup enables you to mirror your production server to a backup server. In a disaster recovery situation, you can easily convert your backup server to be the production server.
This chapter contains instructions for configuring Live Backup for a single server deployment. For multi-tenant deployments, follow the configure Live Backup instructions in the multi-tenant guide.
Server actions are mirrored in real-time. There might be pending actions due to high server load, connectivity issues, etc. Consider the following:
  • Live Backup uses a single main server and a single standby server. Beyond these, additional servers are not currently supported.
  • Active/Active configuration is not currently supported.
  • Each host retains its own distinct IP address and host name.
  • Neither host has any awareness of which host is truly active. Failover is not dynamic, meaning that making a host active must be done manually, by an administrator.
In the event of a server failover, engines dynamically reconnect to the active host.
As the process of making a Cortex XSOAR server active is a manual process, it is conceivable that two servers could be active simultaneously. You must avoid this scenario because both hosts collect and work on potentially the same security incidents, which could possibly lead to the following:
  • Incident duplication
  • A higher load on your integration endpoints
  • Possible significant database inconsistencies due to duplication of internal identifiers being shared between nodes and causing existing incidents to be overwritten.
If there is ever uncertainty about whether a host that is presently down or stopped was in an active state before it went offline, it is recommended that you put the presently active host into a standby state before starting the Cortex XSOAR service on the other host. You can then make it active again after you have confirmed whether the host you are starting is already in active mode.
To configure the live backup environment, see Configure the Live Backup Environment.
The following scenarios describe how to test, and deal with active server failures:
When you first install the Cortex XSOAR server and it starts for the first time, you can use a configuration file to transition between DR states, as described in Transition Between DR States Through the Configuration File.
If you need to upgrade your live backup environment, see Upgrade the Live Backup Environment.
For details about the relationship between engines and disaster recovery, see Cortex XSOAR Engines and Disaster Recovery. For information about host names, DNS, and disaster recovery, see Host Names, DNS, and Disaster Recovery.

Troubleshoot Live Backup

If you receive an out of memory error when live backup is enabled, consider changing the server configurations for disaster recovery.
  1. Select
    Settings
    ABOUT
    Troubleshooting
    Add Server Configuration
    .
  2. Add the following configurations:
    Key
    Description
    Value
    dr.batch.size
    Controls the number of actions sent to the disaster recovery server in one request. A very high value can cause memory issues. A very low value can cause performance issues, which causes the backup server to be synced slower (not in real-time).
    It is recommended to start low (25-50) and increase according to memory usage.
    Default is
    300
    .
    For versions 6.0.1 and below, default is
    500
    .
    dr.memory.limit.mb
    Limits the memory size (in MB) of the action items, which should prevent out of memory errors.
    dr.batch.size
    and
    dr.memory.limit.db
    work together, so the threshold is reached when the limitation of either configuration is met.
    If you receive an out of memory error, consider reducing to 100.
    Default is
    300
    dr.queue.size
    The total number of actions to keep in memory before entering recovery mode. It is recommended to keep the default number, as it is relative to the size of the
    dr.batch.size
    configuration.
    Default is
    10*dr.batch.size

Recommended For You