4.2. Monitoring Node Performance¶
To monitor the performance of a cluster node, open the NODES screen and click the node. On the node overview screen, you will see performance statistics described below.
The overall statistics include:
The number of CPUs and the amount of RAM
CPU usage, in percent over time
RAM usage, in percent over time
The DISKS section shows:
The number of HDD and SSD drives and their statuses
Node I/O activity over time on the read and write charts
The NETWORK section shows:
The list of network interfaces and their statuses
The amount of transmitted (TX) and received (RX) traffic over time
The following sections provide more information on disk and network usage.
4.2.1. Monitoring Node Disks¶
To monitor the usage and status of node disks, click the DISKS link on the node overview screen. You will see a list of all disks on the node and their status icons.
A disk status icon shows the combined status of S.M.A.R.T. and the service corresponding to the disk role. It can be one of the following:
Ok. The disk and service are healthy.
Failed. The service has failed or S.M.A.R.T. reported an error.
Releasing. The service is being released. When the process finishes, the disk status will change to Ok.
On this screen, you can:
Monitor the details and performance of each disk.
Manage disk roles.
Have the disk blink its activity LED. Works only for LSI and PERC controllers.
To monitor performance of a particular disk, select it and click Performance. The Drive performance panel will display the I/O activity of the disk.
To view information about the disk, including its S.M.A.R.T. status, click Details.
To have the disk blink its activity LED, select the disk, and click Blink. To have the disk stop blinking, click Unblink.
22.214.171.124. Monitoring the S.M.A.R.T. Status of Node Disks¶
The S.M.A.R.T. status of all disks is monitored by a tool installed along with Virtuozzo Storage. Run every 10 minutes, the tool polls all disks attached to nodes, including journaling SSDs and system disks, and reports the results to the management node.
For the tool to work, make sure the S.M.A.R.T. functionality is enabled in node’s BIOS.
If a S.M.A.R.T. warning message is shown in the node status, one of that node’s disks is in pre-failure condition and should be replaced. If you continue using the disk, keep in mind that it may fail or cause performance issues.
Pre-failure condition means that at least one of these S.M.A.R.T. counters is not zero:
Reallocated Sector Count
Reallocated Event Count
Current Pending Sector Count
126.96.36.199. Slow Disk and Slow CS Analyzers¶
Slow disk and CS analyzers calculate disk health according to the average I/O latency over time (15 minutes).
The following table shows the default thresholds:
OK latency, in seconds
Fatal latency, in seconds
Slow HDD Disk
Slow SSD Disk
If disk latency is less than OK latency, the disk health is considered to be 100%. If disk latency exceeds FATAL latency, the disk health is considered to be 0%. Disk latency that lies within these two thresholds will vary linearly from 100% to 0%.
When disk health becomes 0%, the service generates an alert and marks this CS as unresponsive. Such a CS does not trigger automatic replication but is no longer available for chunk allocation.