Sun Microsystems, Inc.  Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition
   Home | Current Systems | Former STK Products | EOL Systems | Components | General Info | Search | Feedback

Asset ID: 1-73-1000882.1
Update Date:2010-09-01
Keywords:

Solution Type  FAB (standard) Sure

Solution  1000882.1 :   Sun StorEdge 3310 Arrays with firmware prior to 4.13 and CLI software prior to 2.1, and SAF-TE firmware prior to 1170, may experience downtime, drives offline, and inaccurate status reporting for components.  


Related Items
  • Sun Storage 3310 Array
  •  
Related Categories
  • GCS>Sun Microsystems>Sun FAB>Standard>Mandatory
  •  

PreviouslyPublishedAs
201168


Product
Sun StorageTek 3310 SCSI Array


Impact

Sun StorEdge 3310 Arrays with firmware (FW) versions earlier than 4.13, CLI software earlier than 2.1, and SAF-TE FW earlier than 1170 may experience system downtime and data integrity issues. FW 4.13, available in patch 113722-11, addresses these issues and provides product improvements.

Below are details for some of the major issues addressed by the 4.13 FW:

  • DRAM Parity Errors or SDRAM ECC Errors on Sun StorEdge 3310, Sun StorEdge 3510 or 3511 FC Arrays May Cause File System Integrity Issue.  See Sun Alert 57612.

This issue can occur when the controller firmware fails to distinguish between single-bit ECC errors and multi-bit ECC errors. The controller seems to continue to work normally even for multi-bit errors, which leads to loss in file system integrity. A single-bit ECC error is recoverable, while a multi-bit ECC error is not. With 4.13 FW if this issue happens the controller will shutdown itself.

  • Sun StorEdge 3310/3510/3511 FC Array Controllers May Incorrectly Offline Drives.  See Sun Alert 57702.

During the recovery from a failure, Sun StorEdge 3310/3510/3511 FC array controllers may incorrectly offline good drives causing multiple drive failures. As a result, logical devices may become degraded thereby causing applications to stop running. The 4.13 FW has proper procedure for fault handling and will not cause this issue.

  • Sun StorEdge 3310/3510/3511 Disk Rebuild Operation Fails to Complete.  See Sun Alert 57690.

In the event of a disk failure, disk rebuilding would commence on the spare drive (if configured) and the rebuilding may stop after 99% and not complete. The rebuild will remain incomplete and the logical device state would remain as degraded. Should another drive failure occur, this condition could result in loss of data integrity. This issue has not been seen throughout the extensive tests with 4.13 FW and it is believed the issue is fixed due to the big changes in the fault handling area of the 4.13 FW.

  • Changing the Cache Optimization Mode Incorrectly on Sun StorEdge 3310, 3510, 3511 may cause issues affecting filesystem availability and data integrity.  See Sun Alert 57644.

The 1.6.2 CLI release and subsequent releases prevent the user from changing the cache optimization mode while there is an existing LD. Also, 4.13 controller firmware release has the section of code rewritten so that different mode LDs can exist in a controller, thus making the issue nonexistent.

  • sccli "show frus" command returns inconsistent results intermittently.

The "show frus" command is performed by doing a series of Read Buffer commands that read the FRU data from I2C EEPROMs in the chassis. A FRU would not be displayed when the SAF-TE controller firmware returned a FRU Read Failed sense code. The I2C driver firmware in the SAF-TE controller had several issues that caused I2C messages to be missed. The driver was improved to detect and recover from I2C errors. In addition, message retries were implemented so that failed messages would be recovered. This is resolved in SAF-TE revision shipped with 4.13/2.1 release.

  • For Sun StorEdge 3310, Running SSCS and sccli(1M) In-band at the Same Time May Cause SCSI Errors.  See Sun Alert 57558.

The issue occurs when multiple Send Diag commands are received by the controller. The Send Diag command is single threaded. The firmware does not properly handle returning a BUSY status when a Send Diag is received with a Send Diag already in process. This is caused by the pass through structure being overwritten by the receipt of the second command resulting in inconsistent results including bug hang, bus phase error and/or data returned incorrectly. In most cases a bus reset is issued by the initiator when this issue occurs. This is fixed in FW 4.13.

 

More Info about the 4.13 FW / CLI 2.1 Release

The 4.13/2.1 release is a FW and software upgrade and does not require a hardware change. While the FW updates for this product have been non-disruptive for previous releases, due to the big difference between the current code(s) and this release, the upgrade is a disruptive process and requires a controller reset.

Firmware version 4.13 and CLI software version 2.1 add the following new features to StorEdge 3310 RAID arrays:

1. Common source code for RAID controller firmware with separate bindings specific to FC, SATA, and SCSI.

2. Improves the interoperability with StorADE in regards to:

  • Instrumentation - Discover arrays, gather telemetry data and retrieve event logs
  • Fault Management - Identify FRU faults by applying pre-established thresholds and policies from the instrumentation data
  • Diagnostics - Ability to invoke diagnostic tools in order to isolate to a single failing FRU.

3. New features:

3.1. Cache specific

  • Independent policies for logical drives (LD) user-configurable cache policy per LD; currently the cache policy is per RAID
  • Write-behind cache mode 

3.2. Fault management specific

  • Automatically switch to write-through mode based upon: Low Battery level, AC loss, Fan Failure, Power supply failure, Notification of high temperature in controller or enclosure
  • Enhanced SNMP trap
  • Automatic system shutdown based on critical environmental conditions

3.3. Logical Device and Logical Volume specific

  • Variable stripe size support (4KB -256KB) per LD. Increments will be done by powers of 2 (i.e., 4KB, 8KB, 16KB, &). Currently the stripe size is set per RAID and can be either 32K for random or 256 for sequential.
  • Increase the total number of Terabytes supported per LD to 64TB for sequential and 16TB for random configurations; these numbers are 2TB and 512GB respectively
  • Increase the total number of supported drives per LD to 36
  • Increase the number of LD's supported per controller to 16; currently this number is 8
  • Automatic availability of RAID sets at start of initialization.
  • 16-byte SCSI Command Data Blocks (CDB)s to support >2TB file system 

Resolution

Upgrade the software for the SE3310 Array by installing patch 113722-11. Follow the detailed procedure given in the README file.

Please use this Customer List to identify sites which may be affected.

  • http://sdpsweb.central/FIN_FCO/FIN/FIN101901/Customer_List.sxc

Use this Customer Letter as needed to communicate the issue to customers.

  • http://sdpsweb.central/FIN_FCO/FIN/FIN101901/SPE/Customer_Letter.sxw

Modification History
Date: 19-SEP-2005
  • Posted an updated Customer_Letter.sxw


Previously Published As
101901
Internal Comments


Please reference the following product manuals as needed.




  • Sun StorEdge 3000 Family Installation, Operation, and Service Manual for the Sun StorEdge 3310 SCSI Array, 816-7290




  • Sun StorEdge 3000 Family Best Practices Manual for the Sun StorEdge 3310 SCSI Array, 816-7326




  • Sun StorEdge 3000 Family FRU Installation Guide, 816-7326




  • Sun StorEdge 3000 Family Rack Installation Guide for 2U Arrays, 817-3629




  • Sun StorEdge 3000 Family RAID Firmware 4.0 User s Guide, 817-3711




  • Sun StorEdge 3000 Family Software Installation Manual, 817-3764




  • Sun StorEdge 3000 Family Configuration Service 2.0 User s Guide, 817-3337




  • Sun StorEdge 3000 Family Diagnostic Reporter 2.0 User s Guide, 817-3338




  • Sun StorEdge 3000 Family CLI 2.0 User s Guide, 817-4951




  • Sun StorEdge 3000 Family Safety, Regulatory, and Compliance Manual, 816-7930



 


Related Information
  • URL: http://sdpsweb.central/FIN_FCO/FIN/FIN101901/Customer_List.sxc
    http://sdpsweb.central/FIN_FCO/FIN/FIN101901/SPE/Customer_Letter.sxw
  • Other: Sun Alert: 57612 57702 57644 57558 57690; FIN: I1174-1, I1176-1, I1147-1

Internal Eng Business Unit Group
KE Authors

Internal Eng Responsible Engineer
kevin.l.doan@sun.com

Internal Resolution Patches
113722-11

Internal Kasp FAB Legacy ID
101901

Internal Sun Alert & FAB Admin Info
Critical Category:
Significant Change Date:
Avoidance: Firmware
Responsible Manager: null
Original Admin Info: null

Product_uuid
3db30178-43d7-4d85-8bbe-551c33040f0d|Sun StorageTek 3310 SCSI Array

Attachments
This solution has no attachment
  Copyright © 2011 Sun Microsystems, Inc.  All rights reserved.
 Feedback