6.2.3 Access Path Setting

In a storage system where a Fibre Channel switch is between a server node and storage, this software automatically changes the security settings supported by each device and sets an access path, as shown in the figure below. If no security mechanism is provided on the device side, setting up security may not be possible. However, the security mechanisms of other devices create access paths. If none of the devices constituting these access paths is equipped with a security function, security cannot be set up but the access paths can be accessed from the server node.

Point

You can set the access path concurrently with assigning volumes to the server using the Web Console.
Refer to "Assign ETERNUS Disk Storage System Volumes to Server" in the ETERNUS SF Web Console Guide for details on the operation.

Note

Each storage system has requirements, such as HBA types used, for guaranteeing operation according to the mutual connection verification status. Set each access path so that these requirements are satisfied. For more information about the requirements, contact the respective storage system distributor.

Table 6.1 Security Information That This Software Sets in Device When Setting Access Path
Device Type		Contents Set by This Software
Solaris server node	Agent available/unavailable	No setting
Windows server node	Agent available/unavailable	No setting
Linux server node	Agent available/unavailable	No setting
HP-UX server node	Agent available/unavailable	No setting
AIX server node	Agent available/unavailable	No setting
Other server nodes	Agent unavailable	No setting
Fibre Channel switch	With zoning setting	WWPN one-to-one zoning is set. (However, this zoning is not set if the Fibre Channel switch side already has sufficient port zoning.)
Fibre Channel switch	Without zoning setting	No setting
ETERNUS DX series	With security setting	Setting of HBA WWPN for security
	Host affinity (zone) setting	Setting of HBA WWPN to the affinity group (zone)
	Without security setting	No setting
Other storage/tapes		No setting

6.2.3.1 Preparations

Notes

Only "System Administrator" account can set access paths.
To change the specified AffinityGroup/zone name on the ETERNUS Disk storage system side of a set access path, delete the path and then set it again with a new zone.
Verify that all Fibre Channel switches cascaded to the Fibre Channel switch associated with the access path setting are correctly recognized in the current status. Once an access path is changed because a Fibre Channel switch is temporarily disconnected from the cascade connection, the old cascade connection cannot easily be recovered. (In this event, execute the cfgClear command to delete zoning information for the Fibre Channel temporarily disconnected from the cascade connection. This enables cascade reconnection.)
When setting an access path that is configured in the FC routing functionality LSAN zone in this software, manually set the LSAN zone in the device in advance.
If a fibre channel switch existing on the configured access path is in No Security status, zoning is not done to the fibre channel switch. When configuring zoning to the fibre channel switch, enable "SAN Zoning Setting" on the Change screen for switch information. For the method to display the screen, refer to "Change FC Switch Information" in the ETERNUS SF Web Console Guide.
Security status of fibre channel switch can be viewed on the Detail Information screen. For the method to display the screen, refer to "Display FC Switch" in the ETERNUS SF Web Console Guide.

Checking Storage Settings

Check the following settings on the storage side before setting an access path:

ETERNUS DX series
By setting up the zone (host affinity) function, you can use storage areas efficiently on a storage area network. To use this zone (host affinity) function, you must set up a Affinity Group/zone on the storage side before setting an access path.
ETERNUS DX60/DX60 S2/DX80/DX80 S2/DX90/DX90 S2, DX100 S3/DX200 S3, DX400 S2 series, DX500 S3/DX600 S3 and DX8000 S2 series
If linked affinity groups are used, then the WWN settings of the server nodes to be used must be set in advance.
Use ETERNUS Web GUI to set the WWN settings. When setting it, select a host response whose Host Specific Mode is in HP-UX Mode.

Connected State of FC Cable Between Devices

When setting an access path, an HBA and CA must be physically connected via a Fibre Channel switch.

Device Status

Each HBA must be connected to a Fibre Channel switch and operate as the N port or NL public.
All of a Fibre Channel switch on the way of the access path setting, HBAs, server nodes, CAs and storages should be operating normally.

6.2.3.2 Setting

Set access paths for the Fibre Channel.

Access Path Setting Procedure

Configure the access path environment by installing additional storage and connecting a Fibre Channel cable, etc.

Preprocessing for the device driver definition file
- When ETERNUS Multipath Driver 2.0.1 or higher is used
  No preprocessing needs to be performed.
- Solaris 9 OS
  - This procedure changes the disk array control driver from sd to hddv when the grmpdautocon command is executed when the hot system expansion function (hddvadmM) is enabled for setting a hard disk driver (hereinafter abbreviated as hddv).
    Check whether the hddv configuration information file (hddv.conf) has been created.
    # ls /kernel/drv/hddv.conf
    If the file is not found, execute the following command and create the hddv.conf file:
    # hddvadm -M
    When the file is found, check whether the file content is of class or parent type.
    (Example)
    Example of hddv.conf of the parent type
    name="hddv" parent="scsi" target=16 lun=0;
    Example of hddv.conf of the class type
    name="hddv" class="scsi" target=16 lun=0;
    The class type of file requires no preprocessing. (All hddv.conf files created by the hot system expansion command [hddvadm M] are of the class type.)
    For the parent type of file, follow the instructions below to initialize the non-hot system expansion environment and move into the hot system expansion environment. To directly use the non-hot system expansion environment, follow the procedure for setting non-hot system expansion (hddvamd m).
    1. Execute hddvadm -i.
    2. Confirm that /kernel/drv/hddv.conf is not present.
    3. Delete the entries made for recognizing the disk array device from the sd configuration information file (sd.conf).
    4. Execute the following command and create an hddv configuration information file (hddv.conf):
    # hddvadm -M
  - In the MPLB2.0 or MPLB2.1 environment with non-hot system expansion specified (hddvadm -m) or in MPHD environment, the commands for adding and removing paths are not supported. Accordingly, the multipath configuration must be canceled once. Execute the destroy command of #iompadm to cancel the multipath configuration.
  - In the multipath disk control (MPLB, MPHD, or HDDV) environment with non-hot system expansion is specified (hddvadm -m) or with normal setting (hddvadm-m), execute the following command on the target server node:
    # hddvadm -i
    By executing this hddvadm -i command, hddv.conf file moves to sd.conf and mphd.conf/mplb.conf file is saved. However, in the hot system expansion environment or in the hddvadm -M execution environment, hot system expansion can be performed without executing the hddvadm -i command. Therefore, this command need not be set.
Select the target server node and add an access path by using the Web Console.
[On a Solaris OS server node with multipath disk control (ETERNUS Multipath Driver 2.0.1 or later)] Execute the command below on the target server node to enable the server node to recognize the storage device.
- To automatically build a multipath device, execute the grmpdautoconf command. Refer to the driver manual for details on the command.
```
# /usr/sbin/grmpdautoconf
```
  To automatically build a multipath device that uses the ETERNUS multipath driver and has a path passing through multiple Fibre Channel switches, execute the command after confirming path redundancy and specifying the "-X" option (because the grmpdautoconf command cannot check whether the redundancy of a path passing through multiple Fibre Channel switches is guaranteed).
```
# /usr/sbin/grmpdautoconf -X
```
[On a Solaris OS server node with multipath disk control (MPLB, MPHD, HDDV2.2 or later) and hot system expansion specified (hddvadm -M execution environment)] Execute the following command on the target server node so that the target server node recognizes the storage unit.
- In Solaris 9 OS, hddv.conf is edited if needed (refer to the multipath disk control (MPLB, MPHD, HDDV) various driver manuals). The following command is executed on the server node to fetch a storage device.
```
# update_drv -f hddv
```
With the above setting, the hddv driver can recognize the logical storage unit. If necessary, execute the command that builds an extension path to the multipath instance or the command that configures the multipath instance. Refer to the applicable driver manual for information about the commands (mplbconfig/iompadm add/iompadm new).
[On a Solaris OS server node, in the multipath disk control (MPLB, MPHD, HDDV) environment with non-hot system expansion is specified (hddvadm -m), or in the environment without multiple disk control] Execute the following command on the target server node to enable the server node to recognize the device:
- If the HBA is SE0X7F21F, SE0X7F22F, XSEFC401AF, XSEFC402AF, SG-XPCI1FC-QF2, SG-XPCI2FC-QF2, SG-XPCI1FC-QL2, SG-XPCI2FC-QF2-Z, SG-XPCI1FC-QF4, SG-XPCIE2FC-QF8, SG-XPCIE2FC-EM4 or SG-XPCIE2FC-EM8
  Check the storage Ap_id using the cfgadm -al command, and make the OS recognize the LUN using the cfgadm -c configure Ap_id command. For details about the commands, check the man command.
- Non-MPLB, MPHD, or HDDV environment (including the tape environment)
```
# touch /reconfigure
# /usr/sbin/shutdown -y -i6 -g0
```
- Multipath disk driver control (MPLB, MPHD, HDDV) environment (hddvadm -m)
```
# touch /reconfigure
# /usr/sbin/shutdown -y -i6 -g0
# hddvadm -m (The required part of sd.conf is moved to hddv.conf, and the saved mphd.conf/mplb.conf file is restored.)
# touch /reconfigure
# /usr/sbin/shutdown -y -i6 -g0 (Existing instances for each driver (mplb/mphd/hddv) are configured when the definition file is restored. For new instances, hddv instances are configured.)
```
According to the environment, execute the command that builds an extension path to the multipath instance or the command that configures the multipath instance. Refer to the applicable driver manual for information about the commands (mplbconfig/iompadm add/iompadm new).
[Solaris OS server node] Arrange the files in /dev/rdsk with # disks -C command.
[Windows] Restart the unit so that the device is recognized. Even if the device is automatically recognized by Windows 2003, it is recommended to restart the unit so that the device is recognized again.
[Linux server node] Restart the unit so that the device is recognized.
[HP-UX server node] Restart the unit to execute device recognition. After restart, check the connection of the device and search for the device file (for example, /dev/dsk/c5t1d0) with #ioscan -funC disk command. If the device file does not exist, create the device file by executing #insf-H <H/W pathname > -e command for H/W path in which the device file is not detected.
[AIX server node] Restart the unit to execute device recognition.
[Other server nodes] Perform operations for device recognition.
When the operating system is VMware vSphere 4 or later, reboot the VM host or, by using VMware Infrastructure Client, execute "Rescan" for the HBA that the access path has been set and perform [Reload Conf.] for applicable server node by using this software. Refer to "Access path management" in "Configuration management" in "4.1.7.1 VMware vSphere 4 or Later" for the using method of VMware Infrastructure Client.
Check the access path from this software.
- [For Solaris OS server node]
  /dev/rdisk/cXtXdX is created on the server node side and the access path is displayed.
- [Windows server node, Linux server node, HP-UX server node, or AIX server node]
  A device is generated on the server node side, and the access path is displayed.

Note

Confirmation of the status of server node (HBA)-server node (HBA) and storage device (CA)-storage device (CA) access paths is performed by checking the zoning information for the fibre channel switch device using the Web Console. Deletion is performed by deleting the fibre channel switch device information in the Web Console.

Set the zoning shown below if the ETERNUS VS850 is operating. These are normally set during VS850 installation but can also be set on the Management Server.

Host zone (zone between the host and the VS850)
Storage zone (zone between the VS850 and storage)
I/O group zone (zone for inter-node communication either within a VS850 I/O group or between I/O groups)
Remote zone (zone for communication between different VS850 clusters)

These zonings are normally set during installation of the VS850, but for the storage zone, it is possible to set the switch device's zoning and the storage device's host affinity using this software's access path setting function.
For zones besides the storage zone, set the zoning for the switch device manually.