Chef Infra Server acts as an enterprise hub for configuration data. Chef Infra Server stores cookbooks, the policies that are applied to nodes, and metadata that describes each registered node that is being managed by Chef. Nodes use the Chef Infra Client to ask the Chef Infra Server for configuration details, such as recipes, templates, and file distributions.
This release of Chef Infra Server includes the following key new features and enhancements:
!!! IMPORTANT UPGRADE NOTE !!! It should be noted that this is a major update that requires scheduled downtime due to reindexing. Full release notes are available here.
Chef Infra Server will continue to be a valuable piece of our Chef Infra kit. We are committed to making the solution easier to maintain while we continue to build Chef Automate Infra Server which advances our vision of providing a unified solution (Chef Enterprise Automation Stack) for infrastructure, security, and application automation.
Chef Automate Infra Server is alternative to the traditional Chef Infra Server which is designed to significantly lower operational costs, improve users experiences, eliminate downtime associated with updates and provide better support for high availability clusters. As we continue to build out functionality in Chef Automate Infra Server we are beginning to deprecate features in Chef Infra Server. The end goal being to eventually migrate all customers to Chef Automate Infra Server.
In the June blog “New Chef Automate Functionality and End of Life Impacts” we announced:
The implementation of Elasticsearch is leveraged by both Chef Infra Server and Chef Automate Infra Server. By leveraging Elasticsearch instead of Solr as the search index of Chef Infra Server, Chef Infra Server 14 positions users to transition to the embedded Infra Server in Chef Automate, as described in the June post mentioned above. Some of Chef Infra Server’s largest customers are already using the Elasticsearch-based search indexing and seeing great results.
Chef Infra Server now uses Elasticsearch as its search index which simplifies the indexing pipeline and reduces operational complexity. Solr and RabbitMQ which were used to support the old search indexing pipeline have been removed.
Improved indexing metrics that will improve Chef’s ability to troubleshoot client reported problems related to search indexing. To support this the /_stats API endpoint will now return metrics about search indexing.
Feature x chef-server-ctl now features the check-config command, which runs only the configuration portion of the reconfigure cookbooks and preflight checks.
Oracle Java is now removed and fully replaced with OpenJDK. This removal resolves a number of CVEs in Oracle Java by switching to a newer, compatible build, which does not have the license restrictions that Oracle Java releases after 8u202 have.
Upgrading to Chef Infra Server 14 will require a reindexing operation for most users. We expect this reindexing operation to take an estimated 2 minutes for each 1000 nodes. However, this estimate can be substantially impacted by your server hardware and the complexity of your Chef data. Generally the size of Elasticsearch indices is observed to be 1/10th of the Postgres database size. To be safe we would recommend that the filesystem containing the /var/opt/opscode directory should have at least 20% free space before upgrade.
Learn more about upgrading to Chef Infra Server here.