Workflows for Big Data
Cisco UCS Director Express for Big Data defines a set of workflows in the UCS CPA folder under Orchestration page.
-
UCS CPA Multi-UCSM Hadoop Cluster WF—This workflow is triggered if you use multiple UCSM Hadoop accounts.
-
UCS CPA Multi-UCSM Splunk Cluster WF—This workflow is triggered if you use multiple UCSM Splunk accounts.
-
UCS CPA Single UCSM Server Configuration WF—This workflow is triggered for every UCSM account.
-
UCS CPA Node Bare Metal—This workflow is triggered per node in the cluster.
-
UCS CPA Node Bare Metal with Storage Profile WF—This workflow is triggered using Cisco UCS S3260 storage server having SSD Server Boot Drive under SAS RAID controller to deploy OS.
Note Disk group policies are automatically created in the Cisco UCS Director Express for Big Data using workflows only for Cisco UCS Manager 3.2 or later version.
Table 1. Disk Group Policy Servers Details Disk Group Policy Required?
Disk Policy Allocation Details
Servers having PCH Controller and and Data drive slots with SSD disks
No
Not applicable
SAS controller with dedicated boot drive slots (20x) and Data drive slots with SSD disks
Yes
For OS boot, if you want to manually allocate the SSD slot numbers to RAID 0 or RAID 1, BD_V1_SSM_RAID0 or BD_V1_SSM_RAID1 should be created. Currently dedicated boot drive slots 201 and 202 are supported only by Cisco UCS S3260 Storage Server.
SAS controller with dedicated boot drive slots (20x) with SSD disks and Data drive slots without SSD disks
Yes
For OS boot, if you want to automatically allocate the SSD slot numbers to RAID 0 or RAID 1, Boot_SSD_RAID0 or Boot_SSD_RAID1 should be created.
SAS controller and Data drive slots with HDD only
Yes
For OS boot, if you want to automatically allocate the HDD slot numbers to RAID 0 or RAID 1, Boot_HDD_RAID0 or Boot_HDD_RAID1 should be created.
Note You can create Boot_HDD_RAID0 or Boot_HDD_RAID1 if SSD disks are available in Data drive slots in addition to HDD.
SAS controller and Data drive slots with SDD and HDD disks
Yes
During OS boot, if you want to manually allocate the slot numbers to RAID 1, BD_V1_ANYM_RAID1 should be created.
Note
In some cases, Wait for Complete Association might fail for Cisco UCS 326x storage servers during UCS CPA Node Bare Metal with Storage Profile after a time out. In such cases, association process might have already failed after Bind UCS Service Profile to Template post 90% of FSM with Storage System initialisation error. You need to perform a re-acknowledgement on the Cisco UCS server and resubmit the Bare Metal workflow.
-
UCS CPA Delete Node—This workflow is triggered if you delete a node from the Hadoop cluster.
-
UCS CPA Cloudera Add Live Node—This workflow is triggered if you add a Live Node for Cloudera.
-
UCS CPA MapR Add Live Node—This workflow is triggered if you add a Live Node for MapR.
-
UCS CPA Hortonworks Add Live Node—This workflow is triggered if you add a Live Node for Hortonworks.
-
UCS CPA Add New Disks—This workflow is triggered if you add new disks.
-
UCS CPA Configure NFS Clients and Servers—This workflow is triggered if you configure NFS clients and servers.
-
UCS CPA Enable HTTPS Modify Protocol and Port Number for Hadoop—This workflow is triggered if you modify protocol and port numbers for Hadoop.
-
UCS CPA Enable HTTPS Modify Protocol and Port Number for Splunk—This workflow is triggered if you modify protocol and port numbers for Splunk.
-
UCS CPA Migrate Splunk Cluster to Multi-Site—This workflow is triggered if you want to migrate an existing Splunk cluster to a multi-site Splunk cluster.
-
UCS CPA Remove Disks—This workflow is triggered if you remove disks.
-
UCS CPA Splunk Upgrade—This workflow is triggered if you upgrade the Splunk cluster.
-
UCS CPA Instant Hadoop Cluster WF—This workflow is triggered if you create an instant Hadoop cluster based on the node count and other mandatory inputs such as the IP address range, memory, and number of interfaces. Cisco UCS Director Express for Big Data automatically creates one UCS service profile and a Hadoop cluster profile template at the back end that are required to create an instant Hadoop cluster. This saves you the effort of manually creating a service profile and a cluster profile.
-
UCS CPA Customized Hadoop Cluster WF—This workflow is triggered if you choose to use a specific UCS service profile. A Hadoop cluster profile template with the specified number of nodes to be created in the Hadoop cluster.
-
UCS CPA Disable High Availability WF—This workflow is triggered when you disable high availability.
-
UCS CPA Enable High Availability WF—This workflow is triggered when you enable high availability.
-
UCS CPA Shutdown Big Data Cluster WF—This workflow is triggered when you shut down the Hadoop cluster.
-
UCS CPA Start Big Data Cluster WF—This workflow is triggered when you power up the Hadoop cluster.
-
UCS CPA Splunk Add Live Archival Node—This workflow is triggered when if you add a Live Archival Node for Splunk.
-
UCS CPA Upgrade Cluster WF—This workflow is triggered when you upgrade the Hadoop cluster.
-
UCS CPA Cloudera Add New Service WF—This workflow is triggered when you add a new service for Cloudera.
-
UCS CPA MapR Add New Service WF–This workflow is triggered when you add a new service for MapR.
-
UCS CPA Hortonworks Add New Service WF—This workflow is triggered when you add a new service for Hortonworks.
-
UCS CPA Multi-Node Bare Metal provisioning WF—This workflow is triggered when you deploy baremetal OS account using Cisco UCS server.
Table 2. Disk Group Policy Servers Details Disk Group Policy Required?
Disk Policy Allocation Details
Servers having PCH Controller
No
Disk group policy is not applicable for PCH controllers, for example, Lewisburg SATA Controller. The LUN gets created automatically by the PCH controller based on the controller definition.
Servers having SAS controller
Yes
The disks are configured for RAID 1. You can manually allocate two disks.
Servers having SATA controller
Yes
The disks are configured for RAID 1. You can manually allocate two disks.
-
Bare-metal OS Deployment —This workflow is triggered per server to deploy baremetal OS. UCS CPA Multi-Node Bare Metal provisioning WF will automatically trigger the Bare-metal OS Deployment workflow when you deploy baremetal OS account per server and monitor it.