Our internal R&D team is busy exploring the bounds of our digital world and as our first offering to the general public we have created BlackOut to fill a much needed role in cloud and cluster applications.
What does BlackOut do ::
Following in the footsteps of Henry J Heniz, Distributed Shutdown seeks to “Do a common thing, uncommonly well”. BlackOut empowers system administrators to shut down clusters of computers from anywhere in the world using a simple interface. It’s easy to install, configure, use, and re-use. With the Click of a button tens or hundreds of servers can be shutdown gracefully and in an order that makes sense to the specific cluster.
How does BlackOut do it ::
Ridiculously tiny services are installed on each cluster node. These services watch for commands from authorized users and periodically update both themselves and other cluster members with the most current cluster configuration.
What is a configuration ::
A configuration consists primarily of dependencies between members; e.g. a DNS server or HDFS Primary should wait for nodes that use their services to shutdown before shutting down themselves. Secondly, nodes can be optionally configured to run a command-line script before shutting down. This is equivalent to K-level RC scripts in Linux.
How does BlackOut guarantee all this ::
Successful cloud-computing is based on the principal of failure. Everything fails eventually and your infrastructure must be capable of handling such failures automatically. Feather Information Technology LLC has engineered BlackOut to leverage industry-leading logic and built-in redundancy to propagate authorized commands. Nodes ‘watch’ the rest of the cluster and as failures occur reroute commands over available links. Cluster configurations are automatically propagated and during shutdown nodes enter a ‘high alert, quick response’ mode to ensure the cluster responds to user commands.
Simply put, BlackOut enables administrators to quickly and easily shut down clusters of computers. BlackOut acts as the cluster-level intelligence to orchestrate graceful shutdowns of groups of servers as quickly as possible with as little effort from the administrator as possible.
- Headless design: there is no ‘primary node’ or ‘master server’ to rely on.
- Automated installation
- Simple, easy to understand software configuration
- Node Dependencies :: determine which servers must wait for which other servers before shutting down
- Node Groups :: determine which nodes will be turned of
- Linux, Windows & Mac supported
- Intuitive User Interface, Shutdown, Abort Shutdown, Reconfigure Cluster
- Built-in 30 second shutdown delay
- Abort shutdown at any time
- Update cluster configuration at any time
Have more questions or want a FREE DEMO VERSION? Contact us! This is our first public release of a product and we’re happy to work with every customer to deliver a quality product to fit your needs.
For those of you technically inclined, our video “BlackOut Usage Example” above is composed of 9 nodes uses the following text file configuration:
# ServersA, B, & C are all in a group called “Infrastructure”
ServerA GROUP Infrastructure
ServerB GROUP Infrastructure
ServerC GROUP Infrastructure
# ServersD, E, & F are all in a group called “Data.Storage”
ServerD GROUP Data.Storage
ServerE GROUP Data.Storage
ServerF GROUP Data.Storage
# ServersG, H, & I are all in a group called “Processing.Servers”
ServerG GROUP Processing.Servers
ServerH GROUP Processing.Servers
ServerI GROUP Processing.Servers
# ServerA has cluster-critical functions and must wait for all other nodes to turn off before starting his own shutdown
# Data Storage should wait for processing to stop so D, E, and F must all wait for G, H, and I
# Server G is in charge of Processing and must wait for H and I