330 likes | 485 Views
Functional Overview. Agenda. StackIQ Functional Overview Cluster Management Q&A / Next Steps. Intelligent Infrastructure Automation. Big Data, OpenStack, Red Hat Linux, HPC. One Platform that handles all Bare Metal and Application Management
E N D
Agenda • StackIQ Functional Overview • Cluster Management • Q&A / Next Steps
Intelligent Infrastructure Automation Big Data, OpenStack, Red Hat Linux, HPC • One Platform that handles all Bare Metal and Application Management • No manual scripting and configurations. Less risk with more efficiency • Enterprise Standards and Processes Supported • Over 1 million servers under management AUTOMATION x86 COMMODITIZATION SCALABILITY The only platform with Intelligent Automation at every layer of the Software Stack
StackIQ Benefits StackIQ automates the provisioning and management of Big Data, OpenStack, Linux and HPC clusters from Bare Metal up through the Applications Layer.
Architecture Comparison Without StackIQ NoSQL Scripts Linux, Cloud, Big Data CONFIGURATION, UPDATES / PATCHES OS MGMT., NEW HW, REPLACE HW MANUAL INPUT TO INSTALLERS Hadoop Script Automation Heterogeneous / OS Bare-Metal Provisioning OpenStack Scripts DevOps Linux, Cloud, Big Data StackIQ Roll Dynamic Software Configuration Scripts & Checklists Altered Scripts & Checklists Altered Scripts & Checklists Automation StackIQ GUI/Command-Line/API Manual Scripts & Config Manual Scripts & Config Manual Scripts & Config Server Config Network Config Network Config Disc Config Disc Config Network Config Heterogeneous / OS Bare-Metal Provisioning Automation Server Image Server Image 2 Server Image 3
StackIQ Platform Enterprise Linux Applications Hadoop NoSQL Big Data / Next Gen. EDW Openstack Cloudstack Open Cloud APIx System Visualization, Diagnostics, Repair Networks Package, Config, and Service Management Disks Operating Systems Bare Metal and VM Provisioning Controllers Heterogeneous Hardware
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
StackIQ Platform System Visualization, Diagnostics, Repair Package, Config, and Service Management Bare Metal and VM Provisioning
Customer Success Stories • Multi-Use Customer with Dynamic cluster: 60 nodes Healthcare Hybrid Hadoop and HPC cluster • Existing Linux cluster running on heterogeneous hardware managed by StackIQ • Added an additional 10 Hadoop nodes followed by 53 nodes • Production Node failures requiring automated repair With StackIQ: • Holistic Management • Dynamic Personalization • Intelligent Automation
Customer Success Stories • New Hadoop user planning to scale: 60 nodes 550 nodes Healthcare Communications 550-node Hadoop project at major wireless carrier Hybrid Hadoop and HPC cluster • 55 nodes growing to 4 clusters totaling 550 nodesrunning Hadoop – heterogeneous hardware & software • Evaluated existing general data center management tools and vendor-specific “cluster” management tools and determined they were not sufficient • Small cluster proved node repair and upgrades un-sustainable With StackIQ: • Scalability • Holistic Management • Intelligent Automation
Customer Success Stories • Existing Hadoop user requiring scale, stability and retention: 60 nodes 550 nodes Financial Services Healthcare Communications 1,000 nodes Beat Red Hat Satellite Server & Puppet in Bake Off Hybrid Hadoop and HPC cluster 550-node Hadoop project at major wireless carrier • Expansion to four heterogeneous clusters with a new server type required significant manual effort • A combination of Red Hat Satellite Server, homegrown scripts, and Hadoop specific management tools failed multiple times and negatively impacted the business • Downtime and Degradation = $8 million in annual costs for a business critical cluster With StackIQ: • Flexibility • Intelligent Automation • Stability
StackIQ Cluster Manager
Fully Automated Installation • Assume nothing • StackIQ Cluster Manager installs all the bits (e.g., OS, libraries, application software) and configures all the services (e.g., network, firewall, disks, application services). • Other management tools usually assume all nodes in the cluster have a base OS installed and they are up on the network • Step 1: Install Cluster Manager node • Step 2: Install backend nodes • The Cluster Manager node is ready to install all backend nodes on first boot • Step 3: DONE! • All backend nodes are fully configured on first boot
StackIQ Cluster Manager Screenshots Global View Monitoring • Left hand column shots contexts for viewing and interacting with the cluster (Global, Appliance, Rack, Host) • The Global Monitoring tab is shown here with roll up data for the cluster
StackIQ Cluster Manager Screenshots Asset Management • Cluster-wide device validation and asset inventory
StackIQ Cluster Manager Screenshots Spreadsheet driven host configuration, application definition, and cluster topology • Host Inventory • Attributes • Define Middleware Topology via Spreadsheet • E.g. Where are my Hadoop Nameservers? • E.g. What is my JVM heapsize for Hive? • Global Config • Host-specific Config • System Architect design the spec • Admin can apply the spec to the cluster
StackIQ Cluster Manager Screenshots Global View Alerting • The Alert box at the bottom provides notices to the administrator on system events
StackIQ Cluster Manager Screenshots Disk Management • The Partition tab shows how disk partitioning is handled dynamically for heterogeneous hardware • Optimized RAID controller configuration automated per host/stack
StackIQ Cluster Manager Screenshots Disk Management • The Partition tab shows how disk partitioning is handled dynamically for heterogeneous hardware • Optimized RAID controller configuration automated per host/stack
StackIQ Cluster Manager Screenshots Cluster Validation • The Validation tab allows you to run cluster-wide or sub-cluster tests on Network, Memory, and, Disk performance • A graph shows results with easily identifiable outliers
StackIQ Cluster Manager Screenshots Replace or Reprovision a Server • Replacing or reprovisioning a server is a single click operation
StackIQ Cluster Manager Screenshots • Example: Hadoop • If a Hadoop Roll is chosen, StackIQ will automate installation and configuration of native Hadoop services
StackIQ Cluster Manager Screenshots • Analyze System Utilization • Example • Launch a Hadoop job • View global utilization metrics • View host utilization metrics • For every host in the cluster • Sample every disk, cpu, network for % utilization • Store samples on the local node • Record samples for 5 years • Global Visualization • Interactive (zoom-able) table view • Last 10 minutes of cluster activity • Can be extremely dense • Real-Time Host Visualization • Scrolling graph • Shows utilization for all devices on the host
StackIQ Cluster Manager Screenshots • Analyze System Utilization • Example • Launch a Hadoop job • View global utilization metrics • View host utilization metrics • For every host in the cluster • Sample every disk, cpu, network for % utilization • Store samples on the local node • Record samples for 5 years • Global Visualization • Interactive (zoom-able) table view • Last 10 minutes of cluster activity • Can be extremely dense • Real-Time Host Visualization • Scrolling graph • Shows utilization for all devices on the host
Cluster & Application Provisioning: Defined stack management with the flexibility to integrate and compliment enterprise standards and processes, speeds time-to-value and improves reliability Cluster & Application Configuration: Automating package management and configuration reduces dependencies on scripting and speeds change requests and repairs Support for Heterogeneous Hardware: Maximize Return-On-Infrastructure by making sure nodes are brought into production rapidly and consistently. Support for Multiple Applications: Repurposing of resources allows the enterprise to dynamically respond to the changing and varied requirements of the Lines of Business. Cluster & Application Diagnostics & Repair: Give business users assurance of maximum uptime and maximum performance with given resources Five StackIQ Production Use Cases & their Benefits in Production