Overload Protection

Each server monitors its own CPU load every <Polling Frequency> seconds. If it detects its CPU load exceeding <CPU Usage Threshold>, it finds another server in the same redundancy group with the lowest load. If redundancy is off, or the current server has the lowest load, no swapping occurs.

Once a server is found, cameras are moved starting with the highest camera ID on this server until <Maximum cameras to move> are moved, or 25% of the total cameras on this system, whichever comes first. Cameras that are not Movable (as per Failover Movability setting) are not moved.

 

 CPU overload protection works if you are using a file storage device (for example, NAS) instead of block storage (such as SAN or a local disk) because the pool of storage does not follow the camera when it is moved. If you have central storage (with a NAS), the storage is always accessible from any server.

To set up overload protection:

1.      Select the Move cameras when CPU load is high to enable all other fields.

2.      (Optional) Select the Allow redundant servers to be used check box. If you select this option, automatic failover will not work. (See Failover.)

3.      Use the default settings or change as necessary for the following:

•        CPU Usage Threshold (percent) - The CPU percentage that must be exceeded before cameras are moved.

•        Polling Frequency (seconds) - The number of seconds between checks.

•        Maximum cameras to move - 25% of the server’s current camera count up to this maximum will be moved.

•        Whenever the CPU is high enough that overload protection is required, a message is sent to the Symphony User indicating Redundancy Off, No server available, or Moved x cameras.