Use Solid State Disks to host swap space in order to increase node availability
(Doc ID 1671605.1)
Last updated on AUGUST 04, 2018
Applies to:Oracle Database - Enterprise Edition - Version 126.96.36.199 and later
Information in this document applies to any platform.
Most node evictions (reboot of a node) that are observed in Oracle Clusterware-based clusters are related to memory pressure on the nodes, not necessarily to network issues per se, which is still the second most common reason. This document explains how the use of Solid State Disks (SSDs) for hosting the swap space can mitigate the impact of memory pressure on a node and thereby prevent, at least delay, node starvation before it affects essential node operations.
Delaying or even preventing node starvation is not a solution to the root cause of the problem. Ideally, any type of memory pressure on a node is prevented so that swapping on a node does not occur. However, in case prevention is not successful and swapping is performed by the Operating System (OS), hosting the swap space (Linux: swap partition, Windows: swap file for example) can mitigate the impact that extensive swapping can have on the cluster operations on the affected server(s).
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!