IBM® Open Platform (IOP) 4.3 with Apache Hadoop based on Ambari, includes a new pack of serviceability tools to assist on the deployment of the product, maintenance and diagnosis.
These serviceability features will help customers to facilitate the installation of the product, as well as technical support personnel to identify operations causing downtime or disruption and restore the product into service.
Requirements
- The script should be run from the Ambari server node.
- Passwordless ssh has to be setup from Ambari server node to the client nodes as detailed in Knowledge Center Documentation (Procedure → Step 3 ).
- The script has to be run as root.
Installation
From BigInsights 4.3 servicebaility tools are packaged in a rpm hosted on our official ambari repos. ambari-server rpm has a dependency on the new ambari-serviceability rpm in such a way that serviceability tools will be always installed in the server node and available to be used by support agents. Once the rpm is installed, the content of the rpm will be laid down under /usr/bin/serviceability in the master node of your cluster.
The ambari-servicebility rpm can be also installed independently of ambari-server. This will be the use case for most of customers when they want to use the prechecker tool before installing ambari.
In order to install this rpm independently , ambari.repo needs to be available on /etc/yum.repos.d and they need to run yum install ambari-serviceability command.
Scripts
The serviceability pack includes a set of python scripts covering the following features:
- Pre-installation checks: prechecker.py script reviews pre-installation requirements and recommendations, such as: ssh setup, repo directories, JDK version, python version, unwanted running processes, essential software, SELinux, DNS, unwanted folders, OS version, existing unwanted alternatives, ports …
Health service checks: healthCheck.py script produces the status and the result of the service checks for the running services.
- Version and configuration: versionAndConfig.py script collects output information regarding service versions, stack version, jar file information, RPM information and options & configurations for the installed services.
- Log collection: Log collection can be done in two ways. In a cluster where the service Log Search is installed, logSearch.py can be used to collect logs. Specific log messages generated for a specific service on any particular date can be fetched using this script. Log collection can also be done by using logCollection.py. This script collects the latest logs for all the services.
- Serviceability Main Script: serviceability.py can be used to run any combination of Health service checks, Version and configuration and Log Collection. If log collection is done using the serviceability main script, it checks whether the service Log Search is running and if so, runs logSearch.py. If Log Search is not available it uses logCollect.py that doesn’t need Log Search
For more details on how to consume this set of scripts and what is the expected output, consult the README file available in /usr/bin/serviceability/ folder of the master node of your cluster
In case of product downtime or disruption contact your IBM customer support representatives and they will provide you guidance on how to use these serviceability tools.