Solaris[TM] Errors During Boot with "Failed to create nodes for pwwn"

Asset ID:	1-72-1006168.1
Update Date:	2011-05-09
Keywords:

Solution Type Problem Resolution Sure

Solution 1006168.1 : Solaris[TM] Errors During Boot with "Failed to create nodes for pwwn"

Related Items


Sun Storage 6320 System
 Sun Storage 6540 Array
 Sun Storage 9990V System
 Sun Storage 3510 FC Array
 Sun Storage Flexline 280 Array
 Sun Storage SAN Foundation Software
 Sun Storage 2510 Array
 Sun Storage T3 Array
 Sun Storage T3+ Array
 Sun Storage 6580 Array
 Sun Storage 9970 System
 Sun Storage D280 Disk Subsystem
 Sun Storage 2530 Array
 Sun Storage 2540 Array
 Sun Storage 6020 Array
 Sun Storage 6120 Array
 Sun Storage 6130 Array
 Sun Storage 6140 Array
 Sun Storage A3500 SCSI Array
 Sun Storage D178 Disk Subsystem
 Sun Storage 6780 Array
 Sun Storage 6920 System
 Sun Storage D240 (StorEdge) Disk Subsystem
 Sun Storage Flexline 240 Array
 Solaris SPARC Operating System
 Sun Storage 9960 System
 Sun Storage Flexline 380 Array
 Sun Storage 3511 SATA Array
 Sun Storage 6180 Array
 Sun Storage 9980 System
 Sun Storage 3910 Array

Related Categories


GCS>Sun Microsystems>Storage - Disk>Modular Disk - Other

PreviouslyPublishedAs
208637

Applies to:

Sun Storage 6130 Array
Sun Storage 6140 Array
Sun Storage 6180 Array
Sun Storage 6320 System
Sun Storage 6540 Array
All Platforms
***Checked for relevance on 09-May-2011***

Symptoms

If Fibre Channel devices are not present or turned off, boot errors will occur. Without corrective action, these errors will persist for every re-boot.

The boot errors which are linked to this problem, can be found in the /var/adm/messages file, and are formatted as follows:

luxadm[90]: [ID 593423 user.error] ID[luxadm.create_fabric_device.2316]
configuration failed for line (/devices/pci@8,600000/SUNW,qlc@2,1/fp@0,0:fc::256000c0ffc028e4)
in file: /etc/cfg/fp/fabric_WWN_map. I/O error

scsi: [ID 243001 kern.warning] WARNING: /pci@8,600000/SUNW,qlc@2,1/fp@0,0 (fcp4):
Failed to create nodes for pwwn=266000c0ffe028e4; error=5

Cause

The reason for these errors is that the device in question cannot be seen by the Solaris[TM] Operating System that is connected to a Fibre Channel Storage Area Network. Solaris[TM] will continue to report this error, every time it boots, until corrective action is taken. This error only pertains to fabric (switch) attached devices, as direct attached devices do not invoke the luxadm command at boot that can error as above.

Solution

If device has been moved or removed, execute the command:

# cfgadm -o force_update -c unconfigure <c#::wwn>

where <c#::wwn> represents the controller number and World Wide Number(WWN) of the device that has been removed.

To list all the Fabric devices that should be configured, use the command:

# cfgadm -la

The "c#" can also be found in /dev/cfg by doing `ls -l` on the physical path listed in the error message.
In the above example, the designation would be "c3::256000c0ffc028e4".

To avoid this problem in the future, use the following command before removing the device from the Storage Area Network(SAN):

# cfgadm -c unconfigure <c#::wwn>

If no device has been moved or removed:

forceloading the following drivers during the boot process in /etc/system can resolve the issue.

forceload: drv/qlc
forceload: drv/fctl
forceload: drv/fcp
forceload: drv/fp

If the following drivers are being used by the system it would be worth forceloading these as well.

forceload: drv/ssd
forceload: drv/scsi_vhci

Note: If problem persists, this error is due to a different problem, and further investigation is required.

Attachments

This solution has no attachment