How to check that a low replication speed is caused by the source machine

Challenge

The Acura Control Panel shows that one of the machines is being replicated with an undesirably low speed.

Cause

There are several factors that may interfere with a replication and cause a "bottleneck" issue for its speed. But before focusing on the network bandwidth capacity or potentially faulty settings, please consider that the disk being processed may be solely responsible for a low reading speed due to a poor IOPS performance on the source machine.

In order to gather the relevant statistics and provide their interactive visualization, every Acura instance is provisioned with a Grafana web application. Its interface is available at port 30082 (http:// <Acura's IP:30082>/). Please contact our Support team to get the access credentials. 

Once logged into Grafana, select the " Stats per device" dashboard by clicking on the " Home" drop-down menu in the top left corner and then on the  Default folder.

The Machine ID is available in the Acura UI. Сlick on the green circle with the letter "i" to the left of the machine's name in the Machines Groups section. Then copy the Machine ID to Grafana into the device_id field.

The graphs can be generated according to customer, device or snapshot. A desired time frame can be selected in the top right corner.

The highlighted stats (see the picture above) provide a quick hint towards the source disk performance.

The " agent_stats.Disk Read Speed" value shows how fast the agent is reading from the disk. Its max level can indicate the current replication speed limit..

The " agent_stats.Agent Disk Reading Time" value shows the time needed by the agent to read a block within a single request. The numbers can then be compared to those of healthy machines in order to detect a corresponding low point of disk performance.

Once the issue has been identified, the customer's decision to proceed may vary depending on the migration/DR strategy, which will require further adjustments should time be a pressing concern.

• Related Articles

• How to check that a failed replication is caused by VSS issues

  Cause A failed replication on Windows may be caused by a range of preceding internal issues affecting the source machine. A properly functioning Volume Shadow Copy Service is one of the main prerequisites for a successful replication. Solution Access ...

• How to replace Windows replication agent's config file in case a machine is shown offline in ACP

  Challenge A machine is shown offline in ACP, although the source instance is running and has been replicated successfully before. Cause The replication agent's config file might have become corrupted during an unexpected termination of the repman ...

• Replication fails with a message: Not enough free space in disk to create VSS snapshot

  Challenge  Replication of a machine fails with a similar error message in the Events section: Cause Although the machine might have enough free space in its disks as prescribed in our official Deployment requirements, the actual disk settings can ...

• Linux machine can't mount a disk and boot.

  Challenge A Linux machine can't mount a disk and boot. Cause A wrong disk label is written in fstab, so the machine can't mount the disk and boot. Solution 1. Type in your root password and press "Enter" to log in to a maintenance session. 2. Run ...

• "Module build for the currently running kernel was skipped" message in console during Replication agent's installation

  Challenge During the installation of a Linux replication agent (hlragent) with DKMS, the following message can be displayed in the console: Module build for the currently running kernel was skipped since the kernel source for this kernel does not ...