MiniCluster: Running TFA in the HA/NFS/ACFS/ASM Kernel Zone Grid (Doc ID 2769215.1)

Last updated on JANUARY 06, 2022

Applies to:

Purpose

When this grid has problems the DB/RAC teams will want to have TFA diagnostic collection in order to provide an analysis of any outage and subsequent action plan to resolve.

The way TFA comes deployed 'out of the box' on a new MiniCluster system will only work on one node at a time, while it is normally expected to capture data from all nodes.

That means in MiniCluster KZ Grid one must run it in manually "at the same time" in both nodes to get a complete set of data.

The reason for this is that when it installs it expects to be able to 'scp' and 'ssh' between nodes as user 'root'. Since UID root is a 'role' in MiniCluster, and since tfa is not a 'user' who can assume the role, and since 'passwordless' ssh is disabled by default for user root, TFA needs manual intervention to configure things so that it may run as expected.

This is a 'one-time' procedure which allows one to set up TFA such that it can be run as one command in one node, and will then collect data from each node within the same shell session.

Scope

MiniCluster Kernel Zone HA/NFS/ACFS/ASM Grid is a High Availability ASM/ACFS file system instance that provides shared storage to the rest of the node of the cluster, including global zones and application and database nodes.

The shares are sourced out of a two node cluster based on Solaris Kernel Zone implementation. All of the shares and cluster resources are managed by Oracle Cluster CRS and sourced out of an Oracle Cluster ASM instance in the Kernal Zone Grid.

Details