Use IBMSVC to build a vSphere inter-storage Cluster

Use ibm svc to build a vSphere inter-storage Cluster

Purpose

This document describes how to use ibm svc to build a Vsphere inter-storage cluster.

Solution

What is vMSC?

The vShpere inter-storage cluster is a storage configuration for the VmwarevSpere environment, which has been verified and can be viewed on the specified HCL. All Storage devices that support vShpere Storage clusters can be viewed on the Vmware Storage Compatibility Guide.

What is ibm san Volume Controller?

The ibm svc nodes are configured to pull the remote cluster and use VmwarevSphere for vMotion migration, and the virtual load on the physical data center is switched between the data centers. This solution is briefly presented.

Ibm svc is an enterprise-level storage virtualization system that manages all storage resources. SVC can take over different storage systems, including IBM storage and non-IBM storage, while SVC has common Replication Services, data transmission services, and improves the performance and availability of storage systems. SVC has a unified modular design and is easy to expand. SVC is usually used in a data center to provide corresponding services. SVC can also be configured as a pull-away cluster to provide services for the two data centers (the distance between the two data centers is 300 KM ). The SVC configured in the pull-away cluster provides the pull-away volume, which can be accessed by both data centers at the same time.

What is an ibm svc fixed disk?

SVC's fixed disk is a specific reserved area for cluster management. The SVC cluster has one active quota disk and two standby quota disks. The cluster uses a quota disk for two purposes:

Handle some SVC errors, such as when the faulty node is normal again. Maintain a copy of the cluster configuration information. Each quota disk has a MB zone for this purpose.

Generally, the quota disk of a pull cluster is configured to the third site to prevent the cluster from being affected by the first two sites.

Configuration Requirements

The configured environment must support Vmotion between VMwareHA, DRS, and remote SVC:

VMware vCenter must be able to connect to the vShpere Host IP address of two sites with a minimum bandwidth of 622 Mbps (for migration between vSphere hosts) and a maximum latency of 5 ms vSphere host migration, 10 ms-The IP address of the Enhanced migration source and target vSphere hosts must have The same IP mask and broadcast domain the same IP network on which The virtual machines reside must beaccessible to vSphere hosts at both data center locations. datastores on which the virtual machine boot drives reside must beaccessible to vSphere hosts at both data center locations. the maximum number of vSphere hosts cannot exceed 32 in the HA cluster. The software version of the ibm svc cluster must be later than version 5.1.

Deploy connections between ibm svc Clusters

Ibm svc can be used as a unified vMSC Host channel. The unified Host channel device means that the host can access the local or remote site through the connection between clusters on any site. SVC supports connection between two clusters.

The path between the node and the node does not only pass through the cascade line of the switch. In this configuration, a node is directly connected to the local switch and the remote switch, and the active arbitration disk belongs to two nodes through the optical fiber network. This type of configuration requires the SVC cluster software version to be higher than 5.1.0, and the maximum distance between nodes is 10 KM. The following figure shows the configuration.

The other is that the path between the node and the node goes through the cascade line between the switch. In this configuration, SVC nodes are connected only to the local site, and cascade switches to provide connections to remote switches. The SVC node can access the arbitration disk Through cascade between switches. For this type of configuration, the SVC software version is higher than 6.3.0, and the maximum distance is 300 KM. The following figure shows the configuration.

Applicable scenarios

The following lists the application cases using the remote SVC and VMwarevSphere. The remote SVC can provide a channel to access VMwareVMFS for the vSphere hosts in the two data sources with 300KM data. By pulling a remote SVC cluster, the distance between hosts in the vSphere cluster is 300KM. VMware's vMotion can be performed on hosts with a distance of 300KM. You can perform a virtual machine HA failover between two data centers.

Test scenario

Fault scenarios	SVC response	VMware HA
Path fault-SVC backend Port	A single path between SVC and the storage room fails, and the image between the volumes is not affected.	No effect
Path fault-SVC frontend Port	A single path between the SVC node and the ESXi host fails. ESXi uses an exchange Path Policy.	No effect
Backend storage fault-site1	SVC continues volume copying in site2. When site1 is available, it is synchronized to site1.	No effect
Backend storage fault-site2	SVC continues to copy the volume in site1. When site2 is available, it is synchronized to site2.	No effect
SVC node failure	SVC continues to provide a channel to access the volume through other nodes	No effect
Site1 full fault (faults include ESXi host and SV node)	SVC provides access channels to all volumes through other nodes. After site1 is restored, the image synchronization starts again.	Hosts running on ESXi on the faulty site will be affected. VMware HA will automatically enable the ESXi host on siet2.
Site2 full fault (faults include ESXi host and SV node)	SVC provides access channels to all volumes through other nodes. After site2 recovers, the image synchronization starts again.	Hosts running on ESXi on the faulty site will be affected. VMware HA will automatically enable the ESXi host on siet1.
Multiple ESXi hosts fail-power-off	No effect	VMware HA automatically enables virtual machines on available ESXi hosts.
Multiple ESXi hosts fail-network disconnection	No effect	VMware HA storage for heartbeat No effect
SVC switch cascade faults and vSphere cluster management network faults	SVC uses an arbitration disk to prevent split-brain. One node is specified for activity and the other nodes stop the service.	The ESXi host continues to access the questionnaire through available nodes. No effect
SVC arbitration disk fault	No effect on volume access. A new arbitration disk will be specified, and the volume image will be paused until the new arbitration disk is available and then synchronized.	No effect
Vsphere host Isolation	No effect	HA event dependent upon isolation response rules. Virtual machines can be left on, or rules can dictate for virtual machines to shut down and restart on other hosts in cluster.
Vcenter server faults	No effect	There is no impact on running virtual machines and Vmwware HA. VMware DRS functions will be affected until vCenter is restored.