Enclosures managed by VCM (Virtual Connect Manager) or VCEM (Virtual Connect Enterprise Manager) can be migrated into HPE OneView. Migration recreates the configuration information for an enclosure including hardware, Virtual Connect domain, networks, and server profiles (including MAC addresses and WWNs in the profile connections).
Up to four single enclosure domains managed by VCM can be migrated simultaneously to HPE OneView through the UI or REST API.
VCEM-managed enclosures can be prepared for migration to HPE OneView through the VCEM GUI or a PowerShell module. See Prepare a VCEM enclosure for migration into HPE OneView for more information.
|
|
|
![[NOTE: ]](images/note.gif) |
NOTE: HPE OneView does not support G6 server blades for migration or management. You can replace the G6 server blades with a G7 or later blade, or you can monitor the enclosure. See About monitored c7000 enclosures. |
|
|
When you add an enclosure to be migrated, you can specify an enclosure group or create a new enclosure group. See About managed c7000 enclosures.
More information
You can migrate enclosures into HPE OneView offline or in-service.
|
Offline |
Servers are powered off before migration Advantages: Faster migration time and partially-stacked domain support |
|
In-service |
Servers remain powered on during migration Advantage: Minimum downtime in server connectivity |
Both types of migration require that:
-
All blocking issues, as identified in the compatibility report, are resolved
-
All acknowledgements are understood and accepted, including additional acknowledgments for in-service migration (for example, BIOS and SR-IOV)
In addition, in-service migration requires:
-
A fully stacked enclosure in a dual ring or left/right vertical stacking link configuration. See recommended stacking connections in the HPE Virtual Connect for c-Class BladeSystem Setup and Installation Guide.
|
|
|
|
|
NOTE: Changes made during in-service migration do not take effect until the server is rebooted for the first time following the migration. Plan a reboot if one of the acknowledgements, such as an SR-IOV virtual function configuration, indicated a change which would impact your operation. |
|
|
If the additional in-service migration conditions cannot be met, an offline migration is recommended.
More information
| About blocking issues during migration |
| About migration acknowledgments |
In Virtual Connect Manager, three types of Ethernet stacking configurations are supported and can be migrated into HPE OneView:
|
Every Ethernet interconnect has a path to all other Ethernet interconnects. Full stacking is the default mode. This configuration is migrated into HPE OneView with all Ethernet and Fibre Channel interconnects in one logical interconnect group. This configuration can be migrated in-service or offline. See Full stacking migration from Virtual Connect Manager to HPE OneView. |
|
|
Stacking occurs only between the primary and standby Ethernet interconnects within the domain. This configuration is migrated into HPE OneView with the primary and secondary interconnects configured within one logical interconnect group. All other Ethernet interconnects are configured in their own logical interconnect group with one Ethernet interconnect per logical interconnect group. Fibre Channel interconnects are configured together in a separate logical interconnect group. This configuration must be migrated offline. See Primary-slice stacking migration from Virtual Connect Manager to HPE OneView. |
|
|
Stacking occurs between each horizontal Ethernet interconnect pair. This configuration is migrated into HPE OneView with all horizontally-adjacent interconnects configured into one logical interconnect group. If there is no horizontally-adjacent interconnect present, when migrated, the corresponding logical interconnect group contains only a single Ethernet interconnect. This configuration must be migrated offline. Fibre Channel interconnects are configured together in a separate logical interconnect group. See Horizontal stacking migration from Virtual Connect Manager to HPE OneView. |
You migrate a VCM-managed enclosure into a new or existing enclosure group.
-
New enclosure group — the logical interconnect groups are created based on the Virtual Connect stacking configuration. The networks and uplink sets are copied to the newly created logical interconnect group(s).
For horizontal or primary-slice:
-
Ethernet networks are copied to all logical interconnect groups containing Ethernet interconnects. Uplink sets are copied to all logical interconnect groups. Uplink set ports are copied only to the created uplink set within the logical interconnect groups containing those ports.
All Ethernet settings such as QoS are migrated to the newly configured Ethernet logical interconnect groups.
-
FCoE networks are added to the uplink set created in the logical interconnect group containing the associated uplink set ports.
-
Virtual Connect fabrics are copied as Fibre Channel uplink sets within the logical interconnect group containing the fabric uplink ports. An associated Fibre Channel network is also created.
-
SNMP configurations are migrated to all the logical interconnect groups, including the Fibre Channel logical interconnect group.
-
-
Existing enclosure group — the existing HPE OneView logical interconnect group(s) must match the exact bay and interconnect type of the Virtual Connect configuration. Otherwise, a compatibility error occurs and the migration is blocked.
For example, if a VCM-managed enclosure is Primary-slice and bays 1, 2, 3, and 4 are populated with HPE Virtual Connect FlexFabric 10Gb/24–Port Modules, an error occurs if HPE OneView does not have a matching enclosure group with the logical interconnect groups defined with the same interconnect type (HPE Virtual Connect FlexFabric 10Gb/24–Port Module) for all four bays. In addition, bays 1 and 2 must be configured together in one logical interconnect group and bays 3 and 4 must be configured into two separate logical interconnect groups (as seen in Primary-slice stacking migration from Virtual Connect Manager to HPE OneView).
The attributes applied to the logical interconnect groups such as uplink sets, Ethernet settings, and SNMP configurations must also match or errors can occur.
More information
| About blocking issues during migration |
| Migrate a c7000 VCM enclosure using REST API |
The following VCM settings will not be migrated into HPE OneView.
-
User account specific information such as certificates, user accounts, LDAP, RADIUS, TACACS+, session timeout, and user role configurations
-
Port monitoring configuration. After migration, port monitoring can be configured separately on each logical interconnect. See Configure a port to monitor network traffic in the online help for more information.
-
Address pools of MACs, WWNs, and serial numbers defined in VCM. HPE OneView migrates the MACs, WWNs, and serial numbers that are part of assigned profiles.
NOTE: Any new addresses allocated in HPE OneView after the migration are pulled from the HPE OneView address pools. An administrator can manually define custom ranges in HPE OneView to match those that were defined in VCM to continue allocating from that range.
Before you start the migration process, plan the migration. See “Planning for enclosure migration from VCM into HPE OneView” in the HPE OneView User Guide to make sure certain requirements, such as backing up the VCM configuration, are met.
The migration process is illustrated by the following figure.
See planning your migration in the HPE OneView User Guide for tips on preparing your configuration for migration.
Automated functions
Once the compatibility issues have been addressed and the migration takes place, HPE OneView validates the configuration information and then automatically performs the following functions:
-
Creates a Logical Interconnect Group or uses an existing group.
-
Creates server profiles unassigned to servers but associated to server hardware types. The newly created HPE OneView server profiles retain the same MACs and WWNs used in the corresponding VCM server profiles.
-
Imports the enclosure and adds physical resources such as interconnects and servers.
-
For in-service migration, interconnects are reconfigured sequentially during the migration, overwriting the VCM configuration with the HPE OneView configuration. With a redundant configuration, HPE OneView detects the loss of connectivity for an interconnect and redirects traffic to the alternate path.
At the end of the migration, the Virtual Connect domain is no longer available in VCM.
To migrate an enclosure, see Migrate a c7000 VCM enclosure using REST API.
Before migrating the enclosure, make sure the firmware is at the required minimum. See the HPE OneView Support Matrix for more information. When migrating enclosures, the firmware baseline is automatically set to manage manually. The firmware settings in VCM are migrated to HPE OneView.
HPE OneView does not allow firmware updates during the enclosure migration process.
Once the enclosure is migrated to HPE OneView, you can update the firmware. See Best practices for firmware and Update firmware for enclosures in the UI help for more information.
The following partial list of Virtual Connect features are considered blocking issues because they are not supported in HPE OneView. Blocking issues can also result when settings conflict between the Virtual Connect domain and the HPE OneView logical interconnect group into which the domain is being migrated. HPE OneView checks for these features and blocks the migration when these issues are found. Resolving an issue might require disabling a feature within Virtual Connect, changing a configuration in Virtual Connect, or in some cases, changing the HPE OneView logical interconnect group.
If a feature is required in your environment and your enclosure contains ProLiant G6 or later server blades, you might want to consider monitoring your enclosure. See About monitored c7000 enclosures.
|
|
|
|
|
NOTE: VCEM-managed enclosures can be migrated to HPE OneView. See About migrating c7000 enclosures managed by other management systems for more information. |
|
|
| Blocking issues | ||
|---|---|---|
|
|
|
|
[a] Blocking issue for mixing automatic and non-automatic settings for in-service migration. However, for offline migration, settings are set to default (virtual functions are evenly spread across all physical functions). [b] Blocking issue for in-service migration only. |
||
|
|
|
|
|
NOTE: In general, for functions that are not required in your environment, disabling the function in VCM enables you to migrate the enclosure. |
|
|
An unassigned VCM server profile connection is a connection without a specified network or fabric. There are five types of connections within VCM.
-
Ethernet — One or more Ethernet networks assigned to the same server profile connection.
-
FCoE — An FCoE network associated with an uplink set. Ethernet and FCoE networks may share the same uplink set.
-
FCoE FC SAN — An FCoE connection assigned to a SAN fabric. These connections have one or more native Fibre Channel connections to the uplink ports which are part of the SAN fabric. They utilize FlexFabric physical function 2 of the adapter.
-
Native Fibre Channel — A native Fibre Channel adapter connection to a Fibre Channel interconnect.
-
iSCSI — An iSCSI, Ethernet network or unshared iSCSI network that is associated with an uplink set. Ethernet and iSCSI networks may share the same uplink set.
HPE OneView does not support unassigned server profile connections, so the connections are not migrated and cannot be recreated for the migrated server profile. For native Fibre Channel connections, a warning issue appears on the compatibility report. For all other connections, a blocking issue appears on the compatibility report for in-service migrations.
To determine how to resolve unassigned server profile connections depends on the reason for the unassigned connection and the ramifications to the OS configuration if the connection is not migrated.
More information
| “Unassigned server profile connections cannot be migrated” in the Troubleshooting section of the online help |
| About blocking issues during migration |
| Migrate a c7000 VCM enclosure using REST API |
Some issues that are not considered blocking or warning issues for VCM migration are shown as acknowledgments. Some acknowledgements do relate to warnings in the compatibility report. In either case, acknowledgments are important instructions that must be followed and acknowledged before the migration can begin. Acknowledgments display in the UI after all blocking issues are resolved and a final compatibility report is generated. In the REST API, the acknowledgments are returned in the acknowledgements attribute.
For specific acknowledgements, see the following:
Verify a backup of the VCM configuration (including the output from show config -includepoolinfo) has been obtained and is secured where it will be available in the event that the enclosure migration is unsuccessful. The backup is used if the VCM configuration must be reverted back to Virtual Connect Manager control. The VCM configuration backup can be created through the VCM web UI or the VCMCLI.
More information
| Virtual Connect User Guide at http://www.hpe.com/info/virtualconnect/docs |
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |
Although HPE OneView is fully operational during a migration, do not modify HPE OneView resources until the migration is complete.
|
|
|
![[IMPORTANT: ]](images/important.gif) |
IMPORTANT: Modifying HPE OneView resources can cause the migration to fail, leaving the enclosure unconfigured and requiring recovery before it can be made operational. The migration is complete when the task in the HPE OneView Activities screen shows |
|
|
For example, do not edit networks, network sets, logical enclosures, logical interconnect groups, or server profiles for the enclosures involved in the migration. Do not perform other maintenance activities such as server power, appliance restart, or firmware update during the migration process.
More information
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |
To minimize the impact to applications and services during an HPE OneView in-service migration, redundant hardware and software configurations are required. The fully stacked enclosure must be configured in a dual ring or left/right vertical stacking link configuration. See recommended stacking connections in the HPE Virtual Connect for c-Class BladeSystem Setup and Installation Guide.
Interconnects are reconfigured sequentially during the migration, overwriting the VCM configuration with the HPE OneView configuration. At most, storage is disrupted for a single interconnect at any point during the migration and network traffic is disrupted for the even or odd numbered bays at any point during the migration.
Configure redundant network and storage connections and provide failover such as network interface bonding and MPIO drivers so that network and storage traffic continues to pass on one module of the pair while the other module is reconfigured. As in a logical interconnect firmware update, the server operating system detects the loss of connectivity for the interconnect module and redirects traffic to the alternate path.
More information
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |
In Virtual Connect Manager, administrators can configure specific boot settings (such as PXE or SAN boot) for profile connections while leaving the default USE-BIOS setting in place on other connections in the same profile. After in-service migration, the server continues to boot based on the configured boot settings. However, the first time the server is rebooted after in-service migration, any connection on the profile configured to USE-BIOS is converted to not bootable to match HPE OneView conventions.
More information
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |
In VCM, administrators can specify the distribution of SR-IOV (Single Root I/O Virtualization) virtual functions across the FlexNICs in server profiles. HPE OneView distributes virtual functions evenly across all the FlexNICs in the server profile. When the SR-IOV connection setting in the Virtual Connect domain is default, VCM assigns all SR-IOV virtual functions to the third FlexNIC on the port. When migrating from VCM to HPE OneView, this redistribution takes effect the first time the server is rebooted after the migration to HPE OneView. The redistribution causes the server to reenumerate PCI devices, which can disrupt server network traffic, as a result of the new device ordering. Review and revise the network configuration of hypervisors and guest virtual machines in the host OS, if needed, to avoid disrupting their network connectivity.
More information
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |
Virtual Connect server profiles that are not assigned to server hardware are not migrated to HPE OneView. To migrate these server profiles:
-
Assign them to server hardware, if available.
-
Manually recreate the server profiles in HPE OneView after migration. See Perform post-migration tasks.
More information
| Create a server profile |
| Migrate a c7000 VCM enclosure using REST API |
| About migration acknowledgments |