Understanding Agent Settings (4040379)

The following details Settings are available with Agent and Agentless protected machines. You may use these setting to fine-tune your protected machines.

This KB details the settings available and their expected behaviors.

For the most up-to-date information regarding these settings, see the Rapid Recovery User Guide located at Rapid Recovery Technical Documentation.

Core | Agent | Settings:

1. Open the Core Console, navigate to the Protected Machines menu on the left side of the Core Console, select the machine you wish to adjust, and click on Settings.
2. NOTE: Default values are listed in the Default Setting Values table at bottom of KB. It is recommended that the defaults be used unless environmental conditions determine otherwise.
3. General:
   • Display name: Customizable display name of vCenter / ESX(i) host
   • Hostname: Hostname or IP address of vCenter / ESX(i) host
   • Port: Default ports are changeable if ports are in use or conflict with another program
   • Encryption Key: NOTE: The first time an encryption key is applied a base image will occur. Existing recovery points will remain unencrypted, or if encrypted by their previous encryption keys. When a new encryption key is applied, a base image will be taken, and all-new recovery points will be encrypted with the new key.
   • Repository: Displays repository selected during machine protection. You may only change the repository if the machine has zero recovery points in its existing repository.
4. Nightly Jobs:
   • Depending on the type of protection you have different options. See the Default Setting Values table below.
   • Agent items marked with a Yes* are only available if the Core sees Exchange or SQL server metadata on that Agent.
   • All of these items can be changed on each Agent’s Setting page in addition to the Core’s Settings page.
5. Transfer
   • Priority:
     • Sets the transfer priority between protected machines. Enables you to assign priority by comparison with other protected machines. Select a number from 1 to 10, with 1 being the highest priority.
     • NOTE: Priority is applied to transfers that are in the queue.
   • Maximum Concurrent Streams:
     • Sets the maximum number of TCP links that are sent to the Core to be processed in parallel per agent.
     • NOTE: If you experience dropped packets, try increasing this setting.
   • Maximum Concurrent Writes:
     • Sets the maximum number of simultaneous disk write actions per agent connection.
     • NOTE: For Agent protected machines, Quest recommends setting this to the same value you select for Maximum Concurrent Streams. If you experience packet loss, set slightly lower—for example, if Maximum Current Streams is 8, set this to 7.
   • Use Core Default Maximum Retries: Select this option to use the default retries number for each protected machine, if some of the operations fail to complete.
   • Maximum Segment Size:
     • Specifies the largest amount of data, in bytes, that a computer can receive in a single TCP segment.
     • NOTE: If having Disk issues on Agent or slow transfer speeds, decrease the Maximum Segment Size by half.
   • Maximum Transfer Queue Depth:
     • Specifies the number of commands that can be sent concurrently. You can adjust this to a higher number if your system has a high number of concurrent input/output operations.
     • NOTE: If your system can handle a high number of I/O operations, then increasing this setting may increase transfer speed.
   • Outstanding Reads per Stream:
     • Specifies how many queued read-ahead operations will be stored on the back end (Agent side). This setting helps to control the queuing of agents.
     • DO NOT USE this setting with ESXi Agentless protection.
     • NOTE: If experiencing slow transfer speeds on Agent that has or had highly fragmented volumes, increasing this setting will allow the Agent to read ahead for its next block of data. While there is no maximum value, any setting above 24 rarely improves performance. Increasing this value adds Disk IO and CPU usage on the Agent side.
   • Excluded Writers:
     • Select a writer if you want to exclude it. Since the writers that appear in the list are specific to the machine you are configuring, you will not see all writers in your list.
     • NOTE: Agentless protected machines do not use VSS writers for sending data. All Agent-protected machines have ASR Writer excluded by default. If you include the ASR Writer your backups will fail.
   • Transfer Data Server Port:
     • Sets the port for transfers.
     • NOTE: Default port can be changed if in use or conflict by another program
   • Transfer Timeout:
     • Specifies in minutes and seconds the amount of time to allow a packet to be static without transfer.
     • NOTE: See EXAMPLE below.
   • Snapshot Timeout:
     • Specifies in minutes and seconds the maximum time to wait to take a snapshot.
     • NOTE: See EXAMPLE below.
   • Network Read Timeout:
     • Specifies in minutes and seconds the maximum time to wait for a read connection. If the network read cannot be performed in that time, the operation is retried.
     • NOTE: If your network connection is not established within 10 seconds this implies network issues. Only increase this value in small increments as it has an exponential effect on the total time for a job to fail. It is recommended to extend the Transfer and Snapshot Timeouts first. See EXAMPLE below.
   • Network Write Timeout:
     • Specifies the maximum time in seconds to wait for a write connection. If the network write cannot be performed in that time, the operation is retried.
     • NOTE: If your network connection is not established within 10 seconds this implies network issues. Only increase this value in small increments as it has an exponential effect on the total time for a job to fail. It is recommended to extend the Transfer and Snapshot Timeouts first. See EXAMPLE below.
   • EXAMPLE: The minimum amount of time before a job fails.
     • Assuming streams are not reopened and zero time between network retries. If a stream is reopened the clock is reset.
     • (Network Timeout x Maximum Concurrent Streams) + Job Timeout.
     • Using default values (00:10 x 8) + 10:00 = 11:20. 
6. Excluded Writer:
   • Applies to Agent protected machines only. These settings let you exclude VSS writers. These are machine-specific. A writer is a specific API published by Microsoft to allow other software components to participate in using Microsoft Volume Shadow Services (VSS). Each of the writers in Rapid Recovery that participate in volume snapshots is listed in the Excluded Writers settings. In the event that a writer is interfering with or precluding successful backup transfers, these can be disabled one by one. Quest recommends leaving these settings alone unless you are otherwise directed by a Support representative.
   • ASR Writer is disabled by default. If enabled, the snapshot will fail.
7. License Details: These are details about the license for the specific protected machine. These settings report information from the Core and the Rapid Recovery License Portal. These settings are read-only. To change these settings, update your license information between the Core and the license portal. See your license administrator for details. For more information, see the Data Protection | Rapid Recovery License Portal User Guide.
8. vSphere:
   • Allow Rapid Recovery to delete user-created VMware snapshots:
     • This enables the core to delete user-created snapshots on the host.
     • For example, if you create a snapshot of your host - right-click on the VM, select Snapshot, and click Take Snapshot, this snapshot will be consolidated as part of nightly jobs if selected here and enabled as part of nightly jobs.
   • Allow transfer for volumes with invalid used capacity:
     • For agentless snapshot creation, we use CBT (Changed Block Tracking), a VMware feature. Sometimes CBT will fail due to various reasons. After a failure, the Allocated Space value may be incorrectly displayed at 0% or 100%.
     • If enabled: CBT will refresh automatically allowing the next transfer to process.
     • If disabled: The next transfer will fail until CBT is refreshed manually.