Support > Knowledge Base
Knowledge Base

Search the Knowledge Base: |
Search the Knowledge Base: |
Persistent Binding Issue with Raw Device Mapping
Details
Solution
However, you need to edit the virtual machine configuration files (.vmx) if the virtual machines directly access raw devices and both of the following are true:
- All the LUNs referenced by the virtual machines are still accessible by ESX Server.
- A new cable installation or removal changed the original paths.
For example, assume a LUN originally has four paths:
vmhba3:3:5
vmhba3:4:5
vmhba4:3:5
vmhba4:4:5If a virtual machine configuration file contains lines such as: scsi0:0.present = "TRUE"
scsi0:0.name = "vmhba3:3:5:0"and ESX Server has been rebooted after a pulled cable caused the path vmhba3:3:5 to be removed, then you need to locate the new canonical name for the LUN. If all paths from vmhba3 are no longer accessible, and assuming that target numbers have not changed, then the canonical name will be vmhba4:3:5. In this example, you must edit the virtual machine configuration file and change the scsi0:0.name value as follows: scsi0:0.present = "TRUE"
scsi0:0.name = "vmhba4:3:5:0"If pulled cables, added LUNs and deleted LUNs have caused the SCSI target numbers to change, you need to determine the new vmhba and target number for the LUN and edit the virtual machine configuration file accordingly.
If the virtual machines directly access raw devices or RDMs, and the LUNs referenced by those raw devices or RDMs are no longer accessible by ESX Server, then the virtual machines will not start until you edit the virtual machine configuration file and remove the references to those raw devices or RDMs.
For more information about resetting persistent bindings, see the SAN Configuration Guide at www.vmware.com/pdf/esx25_san_cfg.pdf.
For more information about raw device mapping, see www.vmware.com/pdf/esx25_rawdevicemapping.pdf.
Keywords
Feedback
Actions
- KB Article: 1824
- Updated: Aug 14, 2009
- Products:
VMware ESX - Product Versions:
VMware ESX 2.5.x

