6.2.3. Gear Capacity Planning

Use the max_active_gears parameter in the /etc/openshift/resource_limits.conf file to specify the maximum number of active gears allowed per node. By default, this value is set to 100, but most administrators will need to modify this value over time. Stopped or idled gears do not count toward this limit; a node can have any number of inactive gears, constrained only by storage. However, starting inactive gears after the max_active_gears limit has been reached may exceed the limit, which cannot be prevented or corrected. Reaching the limit exempts the node from future gear placement by the broker.

The safest way to calculate the max_active_gears limit on nodes is to consider the resource most likely to be exhausted first (typically RAM) and divide the amount of available resource by the resource limit per gear. For example, consider a node with 7.5 GB of RAM available and gears constrained to 0.5 GB of RAM:

max_active_gears = 7.5 GB / 0.5 GB = 15 gears

max_active_gears = 7.5 GB / 0.5 GB = 15 gears

Most gears do not consume their entire resource quota, so this conservative limit can leave some resources unused. Most administrators should overcommit at least some of their nodes by allowing more gears than would fit if all gears used all of their resources. Experimentation is recommended to discover optimal settings for your OpenShift Enterprise deployment. Based on the types of cartridges and applications expected, as well as the amount of scarce resources actually used (such as RAM, CPU, network bandwidth, processes, inodes, etc.), determine an overcommit percent by which to increase your limits.

Changing the max_active_gears parameter after installation is harmless. Consider beginning with conservative limits and adjust accordingly after empirical evidence of usage becomes available. It is easier to add more active gears than to move them away.