DataStax Developer Blog

The New OpsCenter 5.0- “Enterprise” Redefined

By Pavan Venkatesh -  July 25, 2014 | 0 Comments

DataStax OpsCenter is a comprehensive monitoring and management tool for managing Cassandra or DataStax Enterprise (DSE). With the new OpsCenter 5.0 which is GA now, we have taken a giant leap by supporting large scale cluster deployments, built-in role based security controls, best practice service, along with other enhancements. You can download OpsCenter 5.0 and start reaping the benefits of the new features.

Certified To Scale Up To 1000 Nodes

OpsCenter 5.0 is certified to support up to 1000 node clusters- on premise or cloud. You can have multiple datacenters each having 100’s of nodes and OpsCenter will be able to visually represent and manage these nodes.

The internal tests have been conducted on Amazon cloud. You can also read “Testing Cassandra 1000 Nodes at a Time” blog written by one of our engineers.

A visual representation of 1000-node Cassandra cluster is shown below.

The diagram represents a condensed ring view- only depicted during large scale cluster deployments with 100s of nodes.

Granular Role Based Security

Basic security (admin, read-only) was provided to the users in prior versions of OpsCenter (4.1.x and below). OpsCenter 5.0 offers advanced, granular based security with user creation and role based permission control, all managed in the same UI (DSE only). This is important for many organizations dealing with compliance and strict internal security requirements.

To enable authentication, edit opscentered.conf file, enable security under the “Authentication” section and restart the opscenterd process.

Click on “Settings” (right top corner of UI) to create users and roles. A default admin account already exists with username “admin” and password “admin”, when authentication is enabled for the first time. We highly recommend to change this default password the first time you log in.

A default “admin” role is created. Only users with the admin role can create users, assign and manage roles for each user, create new clusters etc.

To create a custom role say “user” or “developer”, admin has to click on Manage Roles-> Create Role and a window will appear with various database operations an admin can choose for a role.

For example, users with “Developer” role can only perform certain database operations that have been assigned (as shown in the diagram above). The rest of the operations will be disabled in the UI.

Note that OpsCenter 5.0 uses a new authentication system. The user info (admin, read-only) used in OpsCenter 4.1.x and prior will not work here when you upgrade to the 5.0 version as it points to the old passwd file. Use the script provided in our docs which migrates users and passwords from the old password file and inserts them into the new password database.

Best Practice Service

Often times, a novice Cassandra user may have questions such as: Is my cluster configured properly? Am I following replication and security best practices?  Are my search nodes tuned efficiently with correct settings?

The new “Best Practice” service addresses all these questions. It automatically scans your clusters using a set of predefined rules (20+) and provides expert advice on any found deviations. This helps ensure proper setup and configuration of Cassandra clusters for optimal performance and uptime.

The rules are categorized as advisors and by default all the rules are enabled (‘Best Practice Service’ is available in the services section of the UI).

The performance impact on the cluster is very negligible and we recommend not to disable any rules. You can also configure rules to run at a scheduled time. An indicator is displayed at the top of UI which indicates the rules that have passed or failed. You can click on “Failed” to see the description followed by the expert advice on what to do next.

Let’s walk through one of the failed rules by clicking on it. As shown below, a status message appears saying that SimpleSnitch has been configured and is not recommended in production environments. It also provides best practice advice and recommends user to update the snitch to NetworkTopologyStrategy.

User Interface Restructuring

You’ll see some parts in OpsCenter that have been restructured for simplicity and ease of use.

    • Dashboard: The Performance section has been merged into a single dashboard section.
      • As shown in the diagram, presets now appear as tabs (DEFAULT, DEFAULT1). You can now clone the preset to the same cluster or to a different cluster and also have an option to share it with all users (opscenter authentication should be enabled here). We’ve also added the ability to reorder widgets on the dashboard by dragging the respective title bar.

                             

    • Nodes-> Ring View: The Donut and condensed ring views are the two new ring types that has been added to the UI. The Donut view will be shown only when vnodes are enabled or when ByteOrderedPartitioner is being used. Each node is represented as a slice of a donut based on % of data owned.The new health indicator inside each datacenter ring summarizes health of the nodes (Normal/Medium/High/Down) with a “View in List” option shown  when you click on certain health category (in this case- Normal “3”). This allows you to click on specific nodes and monitor its load, running tasks, TP stats and streams etc.

                         

  • Nodes-> List View: Filters have been added to the various column names so you can directly select, for example, a particular node or group of nodes belonging to analytic DC or even a particular hostname. This is very helpful during large scale deployments.
  • Alerts, editing connection settings for an existing cluster (Cluster Connection), help (Feedback, Report, Diagnostics tarball) have been moved to the top right corner of the UI.
  • Cluster Actions: Cluster wide actions- Add Node, Rebalance, Configure, Restart have been moved from Nodes section to the top right corner of the UI (“Cluster Actions”).
  • Activities Section: This new section displays all running tasks in the cluster. The tasks include actions initiated by either OpsCenter (bulk operations etc.) or by Cassandra (minor compactions etc.). The Event Log, Hadoop Jobs sections, and status when adding nodes to an existing cluster have been moved into this section as well.
  • Data Section: Tables, Data Backups and Explorer sections have been merged into this section.

More information on OpsCenter 5.0 can be found in the docs online.

We’re excited to bring OpsCenter to market for large enterprises. Download OpsCenter 5.0 and start enjoying these benefits. You can reach out to us for any suggestions or enhancements via feedback forum available in the OpsCenter UI.



Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>