Sun Storage 6540 Array
 
Sun Storage 6130 Array
 
Sun Storage 6140 Array
 

GCS>Sun Microsystems>Sun Alert>Criteria Category>Availability
 
GCS>Sun Microsystems>Sun Alert>Release Phase>Resolved
 

1. Impact

In the event of a controller RPA Memory Fault, the array controllers do not log out 
of their Fibre Channel Storage Area Networks(SAN) as expected.

This prevents the scsi_vhci driver from issuing a failover request and 
allows continuous I/O retrys instead of failing them.

2. Contributing Factors

This issue may occur in the following releases:

• Sun StorEdge 6130 Array (on Solaris) without array firmware 06.19.25.16

• Sun StorageTek 6140 Array (on Solaris) without array firmware 06.19.25.16

• Sun StorageTek 6540 Array (on Solaris) without array firmware 06.19.25.16

3. Symptoms

This is an example of an RPA Memory Fault:

   Event Type : 6540.ProblemEvent.REC_RPA_ERR_CTL
   Severity   : 0

   ---- Sample Description ----
   An RPA memory parity error was detected on controller {0}

   ---- Probable Cause ----
   An RPA memory error has been reported on a controller.

   ---- Recommended Action ----
   Replace the controller.


As a result the Array controller will get held in reset, with
the controller tray ID showing "88".  The array logs will start
to fill up with target resets.

   Date/Time: Thu May 22 20:44:02 MSD 2008
   Sequence number: 10330
   Event type: 1202
   Event category: Error
   Priority: Informational
   Description: Fibre channel - TGT reset received
   Event specific codes: 0/0/0
   Component type: Controller
   Component location: Controller in slot B
   Logged by: Controller in slot B

The target resets are the result of the "Retryable" errors from the host(s).
The Solaris /var/adm/messges will fill up with messages similar to:

   May 22 20:02:33 myhost scsi: [ID 107833 kern.warning] WARNING:
   /scsi_vhci/ssd@g600a0b80001111110000121212121212 (ssd54):
   May 22 20:02:33 myhost      Error for Command: read(10)
   Error Level: Retryable
   May 22 20:02:33 myhost scsi: [ID 107833 kern.notice]        Requested
   Block: 113135664                 Error Block: 113135664
   May 22 20:02:33 myhost scsi: [ID 107833 kern.notice]        Vendor:
   STK                                Serial Number:
   May 22 20:02:33 myhost scsi: [ID 107833 kern.notice]        Sense Key:
   Not Ready
   May 22 20:02:33 myhost scsi: [ID 107833 kern.notice]        ASC: 0x4
   (<vendor unique code 0x4>), ASCQ: 0x1, FRU: 0x0

And will lack the "initiating failover" messages typical to controller array 
faults as expected.

4. Workaround

The only workaround is to remove the array connection to the SAN for the
faulted controller, or to offline those ports connected to the faulted 
controller by using a switch management interface for the SAN.

5. Resolution

This issue is addressed in the following releases:

• Sun StorEdge 6130 Array (on Solaris) with firmware 06.19.25.16 or later

• Sun StorageTek 6140 Array (on Solaris) with firmware 06.19.25.16 or later

• Sun StorageTek 6540 Array (on Solaris) with firmware 06.19.25.16 or later

The above firmware is provided by Common Array Manager 6.0.0 or later
releases available at:

http://www.sun.com/download/index.jsp?tab=2


and specifically:

http://www.sun.com/download/products.xml?id=470d094a

This solution has no attachment