Revision history [back]

click to hide/show revision 1
initial version

It can support "create volume from image" with EMC Cinder FC driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

It can support "create volume from image" with EMC Cinder FC driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

We did not deploy Cinder volume service on a VM. You can have a try with the document about NPIV in libvirt

It can support "create volume from image" with EMC Cinder FC driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

We did not deploy Cinder volume service on a VM. You can have a try with the document about NPIV in libvirt

It can support "create volume from image" with EMC Cinder FC driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

We did not deploy Cinder volume service on a VM. You can have a try with the document about NPIV in libvirt. In addition, please feel free to let us know any issue you have.

It can support "create volume from image" with EMC Cinder FC driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

We did not In general, currently there are only two ways to expose vHBA to guest:

  1. NPIV
  2. Host creates vHBA
  3. Use libvirt to expose vHBA as a scsi_host
  4. Single Root Input/Output Virtualization (SR-IOV)
  5. Expose virtual functions of PCI-E devices to guest.

In the case you mentioned, NPIV is not the right solution because vHBAs (NPIV virtual adapters) are SCSI host adapters(and not PCI devices) which are not visible to the userspace and thus are not assignable to guests. This design is in the Linux kernel and cannot be worked around by any userspace component. As far as "How to use SR-IOV", I noticed that there is a document from RedHat (https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Host_Configuration_and_Guest_Installation_Guide/sect-Virtualization_Host_Configuration_and_Guest_Installation_Guide-SR_IOV-How_SR_IOV_Libvirt_Works.html). I hope that it would be helpful. Meanwhile, you need also ensure the FC HBA card (QLE2562) in your environment support the SR-IOV feature.

In addition, we do not have any reference that deploy Cinder volume service cinder-volume on a VM. You can have a try with the document about NPIV in libvirt. In addition, please feel free to let us know any issue you have.

VM with FC Driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

In general, currently there are only two ways to expose vHBA to guest:

  1. NPIV
    • Host creates vHBA
    • Use libvirt to expose vHBA as a scsi_host
  2. Single Root Input/Output Virtualization (SR-IOV)
  3. Expose virtual functions of PCI-E devices to guest.

In the case you mentioned, NPIV is not the right solution because vHBAs (NPIV virtual adapters) are SCSI host adapters(and not PCI devices) which are not visible to the userspace and thus are not assignable to guests. This design is in the Linux kernel and cannot be worked around by any userspace component. As far as "How to use SR-IOV", I noticed that there is a document from RedHat (https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Host_Configuration_and_Guest_Installation_Guide/sect-Virtualization_Host_Configuration_and_Guest_Installation_Guide-SR_IOV-How_SR_IOV_Libvirt_Works.html). I hope that it would be helpful. Meanwhile, you need also ensure the FC HBA card (QLE2562) in your environment support the SR-IOV feature.

In addition, we do not have any reference that deploy cinder-volume on VM with FC Driver.

To access the storage in VNX, Nova compute nodes need to be registered with VNX first. In addition, in order to perform "Copy Image to Volume" and "Copy Volume to Image" operations, nodes running the cinder-volume service (Cinder nodes) must be registered with the VNX as well.

For the detail, you can refer to the session Preparation -> Register with VNX in README of EMC Cinder Direct Driver for OpenStack Juno release.

In general, currently there are only two ways to expose vHBA to guest:

  1. NPIV
    • Host creates vHBA
    • Use libvirt to expose vHBA as a scsi_host
  2. Single Root Input/Output Virtualization (SR-IOV)
    • Expose virtual functions of PCI-E devices to guest.

In the case you mentioned, NPIV is not the right solution because vHBAs (NPIV virtual adapters) are SCSI host adapters(and not PCI devices) which are not visible to the userspace and thus are not assignable to guests. This design is in the Linux kernel and cannot be worked around by any userspace component. As far as "How to use SR-IOV", I noticed that there is a document from RedHat (https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Host_Configuration_and_Guest_Installation_Guide/sect-Virtualization_Host_Configuration_and_Guest_Installation_Guide-SR_IOV-How_SR_IOV_Libvirt_Works.html). I hope that it would be helpful. Meanwhile, you need also ensure the FC HBA card (QLE2562) in your environment support the SR-IOV feature.

In addition, we do not have any reference that deploy cinder-volume on VM with FC Driver.