Understanding Automation: Node Discovery, Inventory, and Quarantine
Intended for use with Cassatt Active Response V5.0.
Automatic node discovery and automatic node inventory are two key time-saving features of Cassatt Active Response. Automatic discovery enables Cassatt Active Response to find any application nodes with integrated power controllers that are cabled into the Cassatt Active Response network by listening for DHCP requests—so it can manage nodes as shared resources. And when you plug in a new node, Cassatt Active Response registers the node in the discovered pool.
Automatic inventory collects a node's hardware characteristics. Cassatt Active Response uses inventory information to determine whether an application node meets the hardware requirements you specify for a tier.
Quarantine is feature that lets you proactively (or on-the-fly) specify nodes that you don't want Cassatt Active Response to manage—so you can use automatic node discovery and inventory with the confidence that if nodes accidentally get discovered, you have a way to isolate them.
This article describe how all three of these features work. You should have a basic understanding of power management as discussed in Understanding Automation: Power Management for Application Nodes.
DHCP and PXE: enabling technologies for node management
Cassatt Active Response leverages two important technologies for automatic node management:
- DCHP – Dynamic Host Configuration Protocol is protocol for assigning dynamic IP addresses to devices on a network. With dynamic addressing, a device can have a different IP address every time it connects to the network. DHCP also supports a mix of static and dynamic IP addresses. DHCP is an extension to the BOOTP protocol.
- PXE – Preboot eXecution Environment is an environment to bootstrap computers using a Network Interface Card (NIC) independent of available data storage devices (like hard disks) or installed operating systems.
In short, Cassatt Active Response uses DHCP for dynamically allocating IP addresses to manage application nodes, and PXE to boot a specific application image on an application node.
How automatic discovery works
Let's first review the requirements for automatic discovery, that is, a full hands-off operation.
Component |
Requirements |
Control node(s)
|
- Control nodes must be the only DHCP server that provides PXE boot services for the networks managed by Cassatt Active Response
- During installation of Cassatt Active Response on the control node(s):
Network mask must match the network mask set in the operating system
Automatic Discovery option must be set to ON (or after installation in the Controller > Discovered Pool > Properties)
You must have a contiguous range of IP addresses set aside for Cassatt Active Response to use to assign to nodes on the Cassatt Active Response network. For nodes that use static IP addresses and are not on a Cassatt Active Response–managed network, the IP addresses must be configured on the remote management interface (or statically assigned using some external DHCP server.)
Power controller user ids and passwords specified during Cassatt Active Response installation, must match those set on the application node, or nodes will not be discovered.
|
Power controller |
Must have a Cassatt Active Response–supported integrated power controller and firmware version (Dell RAC, Dell DRAC, HP iLO, IBM BCMM, IBM RSA II) OR a locally-qualified firmware version of a supported integrated power controller that passes an "on-the-fly evaluation" during discovery. |
Application nodes |
PXE must be enabled in the machine BIOS for HP iLO, Dell RAC, Dell DRAC; IBM RSA II and BCMM set PXE through their web interfaces. |
Steps in automatic discovery
What happens during automatic discovery? More importantly, how does Cassatt Active Response acquire control of an application node for use as a resource? How does Cassatt Active Response leverage power controllers and PXE booting to achieve a hands-off discovery?
The following example is a high-level look at how Cassatt Active Response uses integrated power controllers (Dell RAC), DHCP, and PXE booting technologies for automatic node discovery.
- During Cassatt Active Response installation, you reserve network addresses for Cassatt Active Response.
- In the Controller, you turn on automatic discovery: Discovered Pool > Properties.
- You connect application nodes to the Cassatt Active Response network switches.
- You configure the Dell RAC power controller for DHCP.
- You configure application nodes to PXE boot in the BIOS.
- In preparation for discovery, you make sure application nodes are connected to power but turned off, network switches are connected and turned on, and control nodes are up and running.
- Cassatt Active Response detects a DHCP request from the Dell RAC power controller and validates that it's a Cassatt Active Response–supported power controller with a firmware version it can manage.
- Cassatt Active Response issues an IP address to the Dell RAC power controller.
- Cassatt Active Response uses the power controller to power on the application node and look for a bootable NIC MAC address.
- In response, the power controller powers on the application node and finds a bootable NIC MAC address.
- In the Controller, the application node shows up as a node in the Discovered Pool > Node List.
In the Node Properties page, the node is discovered, but hardware characteristics have not been found.
The Dell RAC power controller shows up as a power controller in the Discovered Pool > Power Controllers.
- As Cassatt Active Response intercepts other DHCP requests, more application nodes get discovered as previously described.
- Application nodes in the Discovered pool are then booted (via PXE) so Cassatt Active Response can inventory their hardware characteristics.
How to use automatic discovery to find slots for blades
For devices like IBM Blade Center BCMM controllers, you may know the BCMM controller MAC address, but not the MAC addresses of blades within the Blade Center rack. In this case, you can configure Cassatt Active Response for automatic discovery by entering the BCMM's MAC address; Cassatt Active Response finds the MAC addresses and slots of the blades. After discovering this information, you can export it to a .cvs file and optionally update the node/power controller information in the Controller.
How automatic inventory works
If you have enabled automatic node inventory (Discovered Pool > Properties), Cassatt Active Response determines the number and type of CPUs, RAM, number of disks, whether an HBA is present, number of NICs, and interface characteristics. Cassatt Active Response uses inventory information to determine whether an application node meets a tier's hardware requirements. Once inventoried, Cassatt Active Response tags the nodes for the free pool and puts them under management control.
As shown in the following diagram, after node inventory, hardware characteristics are populated in the Node Properties page.
After nodes inventory: move to the free pool
If you have turned on the setting, Discovered Pool > Properties > Automatic Move to Free Pool, inventoried nodes will automatically move to the free pool where they are available for use in tiers.
top
Automatic inventory off: when and why?
There are a few circumstances where you might want to turn off automatic inventory:
- To inventory nodes in a careful or serial fashion—as an initial validation of how Cassatt Active Response works
- To make sure you haven't plugged in a wrong node.
- You do not know the MAC addresses of nodes; so you want to bring in one node at a time and enter rack/slot information.
To inventory nodes one at a time, turn off Automatic Inventory, and go to the Discovered Pool and select the option on a node to inventory the node.
FAQ: automatic discovery and inventory
Now that you know the basics, here's some straggling questions you may have.
Question |
Answer |
What happens when Automatic Discovery is turned On? Off?
|
If Automatic Discovery = On, Cassatt Active Response automatically discovers all nodes by listening for DHCP requests. DHCP requests can be detected either from power controllers, or via the PXE setting in the BIOS. Any nodes you have proactively flagged in the Controller as "quarantined devices" are not discovered.
If Automatic Discovery = Off, Cassatt Active Response ignores DCHP requests from power controllers or a node's PXE setting, unless you have manually added them in the Controller using Add Device. |
Can nodes not configured for DCHP and PXE reside on the Cassatt Active Response network? |
Yes. But make sure their IP addresses are reserved from the Cassatt Active Response IP address range so Cassatt Active Response doesn't allocate them to other machines. |
If an application node's power controller is configured for DHCP, but the node is not configured for PXE booting, does it show up in the Controller? |
Yes. You'll see the power controller listed in: Discovered Pool > Power Controllers, but the node will not be listed in the Discovered Pool > Node List (and therefore, will not be managed by Cassatt Active Response.) The exception is the IBM BCMM controller, where Cassatt Active Response can get the MAC address without PXE booting. |
Where do I set Automatic Discovery, and what is the default? |
You configure the setting when you install Cassatt Active Response (ccinstall); you can change the setting in the Controller on the Discovered Pool > Properties page.
Default = Automatic Discovery, ON (in both interfaces) |
What if the power controller login/password needs to be changed prior to discovery? |
Got to: Discovered Pool > Power Controllers. Click the appropriate MAC Address link. |
Can I turn automatic discovery on and off at any time? |
Yes. You initially configure the setting when you install Cassatt Active Response, and can change it any time during in Discovered Pool > Properties. |
Can I manually add nodes/power controllers to Cassatt Active Response at any time, independent of the automatic discovery setting? |
Yes. You can manually add nodes/power controllers at any time, regardless if the automatic discovery setting is ON or OFF. |
Will Cassatt Active Response auto discover blades in an IBM Blade Center BCMM? |
Yes. Blades can be auto-discovered, but Cassatt Active Response gets the information from the BCMM. |
How long does it typically take for Cassatt Active Response to discover and inventory nodes? |
It depends on the type of node and how long it takes to boot. In general, if a node hasn't discovered and shown up in Discovered Pool > Power Controllers within 7-8 minutes, you should probably start troubleshooting procedures. |
Can virtual machines (VMM) be automatically discovered in Cassatt Active Response? |
VMMs are always discovered when their host boots, regardless of the automatic discovery setting. |
top
How to ensure nodes aren't discovered by mistake
You can avoid having nodes discovered that you don't want Cassatt Active Response to manage by following these guidelines:
- Although not required, Cassatt recommends putting all Cassatt Active Response–participating nodes on a VLAN that segregates broadcast traffic (DHCP). This avoids potential conflicts with an existing PXE or a DHCP server.
- Quarantine nodes (see next section)
top
How quarantine works
The quarantine feature allows you isolate nodes that you don't want Cassatt Active Response to manage. You can use the quarantine feature proactively to prohibit nodes from being discovered, or after discovery.
Example: using quarantine to proactively isolate nodes
When you are getting Cassatt Active Response up and running, you may not know all of the nodes that Cassatt Active Response may discover and try to manage. The following steps show how to find nodes that may be issuing DHCP requests that you can then quarantine.
- During Cassatt Active Response installation (ccinstall), set Automatic Discovery = OFF.
- After you've completed configuring Cassatt Active Response, power off the power controller.
- In the Controller, go to Discovered Pool > Properties page and set Automatic Discovery = ON, set Automatic Node Inventory = OFF, and set Automatic Move to Free Pool = OFF.
- Power on the power controller.
Cassatt Active Response discovers all nodes whose power controllers are making DHCP requests; nodes are listed in the Discovered Pool > Node List and power controllers are listed in Discovered Pool > Power Controllers.
- On the Discovered Pool > Node List, select the node(s) that were inadvertently discovered that you don't want Cassatt Active Response to manage.
- From the Power Actions menu, select Delete Node.
In the Delete dialog, you are asked if you want to quarantine the node. The associated power controller is also deleted.
FAQ: Quarantine
The following table provides further information on quarantining nodes/power controllers.
Question |
Answer |
How do I quarantine a node/power controller that has been inadvertently discovered?
|
Discovered Pool > Nodes List > Node Actions > Delete. The delete dialog will ask you if you to quarantine the device so Cassatt Active Response will ignore any further DHCP requests from it.
Important: If the node/power controller you are quarantining has more than one MAC address, and the node has not been inventoried, you must quarantine each MAC address to prevent Cassatt Active Response from discovering the node/power controller. If the node has been through inventory and you quarantine it, Cassatt Active Response will quarantine all the MAC addresses. |
How do I delete a node/power controller that was erroneously added? |
If the node/power controller has not yet been powered on, you can delete it from Discovered Pool > Power Controllers, otherwise, follow the procedure above to delete the node. |
When a node/power controller is quarantined, is it ignored, regardless of the Automatic Discovery setting? |
Yes. Quarantined devices are ignored regardless if Automatic Discovery is ON or OFF.
|
Can I quarantine nodes/power controllers not on the Cassatt Active Response network? |
Yes. Enter the IP addresses of the nodes instead of the MAC addresses. |
top
Discovery and inventory problems
As a guideline, nodes should discover within 7-8 minutes. For nodes with integrated power controllers, see Troubleshooting.
Was this article useful? Tell us what you think.
Email infocentral@cassatt.com.
|
