solaris识别新加存储盘

The technique used will depend upon the type of storage device being added.  This doc may be used as a quick reference guide, however, you should look at the detailed procedures which are available for the specific server, (or storage) that is being allocated.

For any scsi attached jbod device, you can recreate the device trees "on the fly" by reloading the associated driver and re-creating the device trees. In this example, we re-create the device information for a "sd" managed SCSI disk drive in a JBOD array.

• # devfsadm -i sd
• # devfsadm -Cv

Some of the Internal disks in servers use the cfgadm utility to facilitate a disk replacment. Here is an example of replacing an internal failed disk.

• # cfgadm -c unconfigure c1::dsk/c1t3d0 (used prior during removal)
• # cfgadm -c configure  c1::dsk/c1t3d0

Some SAS connected disk drives are hot swapable via the mpt driver.

• # devfsadm -i mpt
• # devfsadm -Cv

Even iscsi uses the same technique.

• # devfsadm -i iscsi
• # devfsadm -Cv

Fiber Channel fabric attached devices use the cfgadm utility. In this example, lun 1 of fabric device 203400a0b82fbc5d is added to the server.

•  # cfgadm -al -o show_FCP_dev
      c2::203400a0b82fbc5d,1         disk         connected    unconfigured   unknown
• # cfgadm -c configure c2::203400a0b82fbc5d,1 

Direct attached fiber channel storage uses the luxadm utility. Some hardware specific commands may include.

• # luxadm insert_device xxxxx
• # luxadm -e force_lip  /xxxxxx
• # devfsadm -i ssd
• # devfsadm -Cv

The common theme to getting all storage dynamically allocated to the server is with the devfsadm command. If devfsadm does not create a device for you (in the formatutility), then its best to consult the hardware configuration and service manual for your particular server.