Sun Microsystems, Inc.
spacer spacer
spacer   Sun System Handbook Home | Systems | Components | General Info | Search | Feedback 
spacer
black dot
 
black fade
 

Sun StorEdge[tm] A5000 Array: Field Information Notices

Implementation (Impl.):
M
= Mandatory (Fully Pro-Active)

C
= Controlled Pro-Active (per Sun Geo Plan)

R
= Reactive (As Required)

FIN
Impl.
Components Affected Description
I0777-1
R
SBus Dual FC Network Adapter (375-3048-01)
SBus I/O Board with SOC+ (501-4266-08)
I/O Board with SOC+ (501-4833-01)
Graphics I/O Board with SOC+ (501-4884-07)
FC-AL SBus Card (FC100/S) (501-5202-03)
FC-AL SBus Card (FC100/S) (501-5266-04)
A3500FC FC-AL Array Ctrlr w/Memory (540-4026-01)
In some configurations, after a Fibre Channel (FC) loop disruption such as a system reboot or a plugging/unplugging of the FC cable at the FC hub, the SOC+ HBA might intermittently take the FC loop down. Also, it is possible that a LIP (Loop Initialization Primitive) received on an idle port of a SOC+ HBA can disrupt I/O activity on the second port of that HBA.
I0712-1
R
FC-AL GBIC Module 100MB/s (option X6731A, 370-2303) This FIN is to inform the field of a new style GBIC that will be shipping with A5X00 Storage Arrays. Installing a FINISAR GBIC in any platform other than A5X00 will result in a restricted access condition which may result in unanticipated downtime.
I0711-1
R
16-PORT Python Switch (370-4276-01) Some StorEdge Network FC Switch-16 units were assembled and shipped from the vendor with their mounting brackets reversed. This could potentially result in these switches being installed backwards in the 72-inch StorEdge Expansion Cabinet. This improper installation may cause the inlet temperature of the switch to exceed its recommended range and this could affect the performance and the life of the switch. In addition, the connecting fibre cables might be damaged by hitting the rear cabinet door and may require replacement.
I0701-1
R
Instant Image 2.0
Fast Write Cache 2.0
The Fast Write Cache 2.0 and Instant Image 2.0 states that the products will work with Java versions 1.2.1 and higher. This is not the case. For installations without Component Manager 2.1, Java versions 1.2.1_04a through 1.2.2_05a are the only versions of Java that will work with Instant Image 2.0 and Fast Write Cache 2.0. If Component Manager 2.1 is installed, then Java 1.2.2 is the only version of Java that will work. Java 1.3 does not work at all with these products.
I0653-1
R
RSM array shelves (370-2343) RSM Array shelves may collapse under weights greater than 40lbs (18.2kg), causing system damage or personal injury.
I0647-1
R
FC-AL SBus Card (FC100/S 501-3060-06)
FC-AL SBus Card (FC100/S 501-5202-03)
FC-AL SBus Card (FC100/S 501-5266-04)

Fcode upgrade instructions for the FC100/S Host Adapter supplied with the following luxadm patches can lead to obsolete fcode being installed.

  • 105310-14 SunOS 5.5.1: Patch for socal, sf driver, and luxadm
  • 105375-25 SunOS 5.6: sf & socal driver patch
  • 107473-06 SunOS 5.7: luxadm patch
  • 109529-03 SunOS 5.8: luxadm, liba5k and libg_fc patch

The latest fcode for the FC100/S Host Adapter has been moved to patch 109400-03 for all versions of Solaris. This change is not reflected in the README files of the luxadm patches currently available (105310, 105375, 107473 and 109529). These patches still provide a lower version of the FC100/S fcode (1.11) and give instructions for installing it.

I0630-1
R
A5x00 Arrays
T3 Arrays
Fiber channel devices may experience symptoms consistent with marginal signal quality on the GBaud FC-AL loops. Marginal signal quality can be caused by dust particles and/or errant electrostatic discharge (ESD) for GBIC's, backplanes, interface boards, interconnect assemblies, FC-AL drives, and fiber cables.
Symptoms of errors that might be seen include:
excessive "CRC Error" or "scsi parity error" messages
excessive "OFFLINE" messages
excessive "timeout" messages
"Offline Timeout" messages
I0575-1
R
Sun StorEdge Component Manager 2.0 (798-2832-01) When used on a heavily loaded system to monitor multiple enclosures, Component Manager response time can be longer than expected.
I0554-1
R
Veritas Volume Manager for Solaris Version 3.0.2 (798-2205-01) There is confusion with regard to supporting the A5x00 products with hubs attached and the use of Volume Manager version 3.0.2. This confusion came from the Veritas Volume Manager Hardware Application Note Release 3.0.2, which states on page 12...

"Dynamic Multi Pathing (DMP) does not support hub-attached A5000,
A5100, and A5200 arrays at this time. See your sales representatives
for details about DMP support plans for this configuration."
If the above recommendation is not followed, then the system could encounter catastrophic results.
I0541-2
R
FC100/P FC-AL PCI Card (375-0040)

Any FC-AL PCI loop where A5X00 Storage Arrays are directly connected to i) a host using PCI FC100/P initiator card without hub, or ii) other host (dual host) with FC-AL initiator-card is connected to the same loop may experience the following errors as described in BugId 4243157.
Sample error messages:

Dec 8 05:18:17 name unix: transport rejected (-2)
Dec 8 05:18:17 name unix:
Dec 8 05:18:17 name unix: WARNING: /pci@4,4000/SUNW,
ifp@4/ssd@w2100002037141659,0 (ssd32)
Dec 8 05:18:17 name unix: transport rejected (-2)
Dec 8 05:18:17 name unix:
Dec 8 05:18:18 name unix: ifp0: LIP occurred
Dec 8 05:18:18 name unix: ifp0: Loop up
Dec 8 05:18:18 name unix: ifp0: Loop reconfigure in progress
Dec 8 05:18:18 name unix: ifp0: Loop reconfigure done
Dec 8 08:17:41 name unix: ifp0: Loop down
Dec 8 08:17:46 name unix: ifp0: LIP reset occurred

UPDATE FOR FIN I0541-2:

The Corrective Action on FIN I0541-1 provides two different options or solutions to the above mentioned problems as follows:

  • OPTION-1: The FC100/P card is the only Initiator on any one loop where Sun A5x00 Storage Array are direct connected.
  • OPTION-2: A Sun FCAL hub is placed between the FC100/P card(s), and the A5X00 Storage Array. In this case, more than one host-adapter on the same loop, may be used.

Due to resistance from the Field over the cost of implementation, OPTION-2 has been removed from the Corrective Action. This option would have required the customer to purchase and connect a hub to the A5x00 Storage Array. The Corrective Action now recommends only OPTION-1 for avoiding this FC100/P card problem. This restriction is advantageous from high-availability/RAS point of view since having each host on a separate loop removes a single point Of failure (SPOF) in Multi Initiator configurations.

I0506-1
R
72" rack There is confusion in the field about a 2" clearance requirement on the left side only (viewing from the front of the cabinet) of the 72" Storage Expansion Cabinet. This clearance is not required for all cabinet configurations, only those containing A5x00 Storage Arrays and/or E4x00 Enterprise Servers.
I0503-1
C
Seagate ST19171FC 9GB Disk Various applications running on Sun StorEdge A5000 units with 9GB ST19171FC drives may report SCSI transport, SCSI reset or read errors. The problem has only been reported thus far with database programs, which may report Assertion Failed errors as documented in BugId# 4187073 and as described below:

Sample Sybase 694 Error: The actual error encountered will vary slightly.

Oct 26 11:13:44 brazil unix: sf4: Target 0x33 Reset successful
Oct 26 11:13:44 brazil unix: WARNING:
/sbus@48,0/SUNW,socal@1,0/sf@0,0/ssd@w2100002037078097,0 (ssd28):
Oct 26 11:13:44 brazil unix: SCSI transport failed: reason 'timeout':
retrying command
/sbus@48,0/SUNW,socal@1,0/sf@0,0/ssd@w2100002037078097,0 (ssd28):
Oct 26 11:13:44 brazil unix: SCSI transport failed: reason 'reset': retrying
command
00:98/10/26 11:13:45.13 server Starting diagnostics for read failure:
00:98/10/26 11:13:45.13 server Device Information:
00:98/10/26 11:13:45.13 server Buffer Information:
00:98/10/26 11:13:45.13 server Page Information from first read attempt:
00:98/10/26 11:13:45.13 server Page Information from second read attempt:
00:98/10/26 11:13:45.15 server SDES Information:
00:98/10/26 11:13:45.15 server PSS Information:
00:98/10/26 11:13:45.15 server End diagnostics for read failure:
00:98/10/26 11:13:45.15 server Suspect a Device or OS problem, Second IO
successful
00:98/10/26 11:13:45.17 server Error: 694, Severity: 24, State: 1
00:98/10/26 11:13:45.17 server An attempt was made to read logical page
'1115947', virtpage '50399017' from virtual device '3' for object '1104006964'
in database '32'. The page was not read successfully. You may have a device
problem or an operating system problem.
I0502-1
R
370-2303-01 GBIC
370-2303-03 GBIC
370-3975-01 GBIC
The Factory phased in shipments of -03 GBICS beginning early April 1999. A5X00s with chassis s/n greater than 9917 will have the -03 GBIC, or an IBM -02 CD GBIC. These modules appear physically identical to the -02 modules. Field personnel cannot differentiate between module revisions without physically removing them to read the revision level on the connector label. This would require an unacceptable maintenance downtime for host systems.The Vixel -03 Vertical cavity GBICS can be identified as follows:
  • Parts are marked as a -03; this marking can be found on the GBIC connector (system interface connector) as "03a" for -03 revision a.
  • The inside if the cable connector housing will be silver in color on the Vixel vertical cavity (-03) parts; black on the -02 cd laser versions.
I0501-1
R
7 slot backplane 501-4372 (-02, -03, & -04)
7 slot backplane 501-5316-04
The revision level printed on the labels for the 7-slot Backplane and the 11-slot Backplane for the A5000, A5100, and A5200 array are different than the revision levels that are dynamically reported by the arrays.
I0465-1
C
GBICs, backplanes, interface boards, interconnect assemblies, or disk drives As of November 1998, approximately 1600 Terabytes of A5000's have been shipped worldwide. A small percentage of these units are demonstrating symptoms consistent with marginal signal quality on the GBaud FC-AL loops. Marginal signal quality can be caused by variations in component quality for GBICs, backplanes, interface boards, interconnect assemblies, drives. Note that the majority of Sun A5000 customers are not experiencing these problems.
Symptoms of this problem include the following:
  • Excessive "CRC Error" or "SCSI parity error" messages
  • Excessive "OFFLINE" messages
  • Excessive "timeout" messages
  • "Offline Timeout" messages
I0421-1
R
Seagate ST19171FC 9GB Disk Recovering from a root filesystem disk failure or corruption on an internal FC-AL boot disk in an E3500 Server or a FC-AL boot disk in a A5000 which is not mirrored with SEVM or SDS.

Note: This FIN was written before there was an official fix for this; also refer to SRDB 17643 if this problem is experienced.
I0418-1
R
Sun Enterprise Volume Manager versions 2.5 and 2.6 The Sun Enterprise Volume Manager (SEVM) vxinstall script may fail to detect a StorEdge A5000(s) attached to a system and prompts for a license number to be entered. This should not be the case since an automatic license should be generated for StorEdge A5000(s). This problem is not reproducible at will although, it has been experienced at several customer sites and in engineering test labs.
I0400-2
R
Seagate ST29102FC (9.1GB 10000 RPM FC-AL 540-3869)
Seagate ST39103FC (9.1GB 10000 RPM FC-AL 540-3869)
Seagate ST318203FC (18.2GB 10000 RPM FC-AL 540-4191)
Empty backplane/incorrect configuration on StorEdge A5x00 can Cause Errors. This minimum configuration is a requirement because circuitry on the disk drives regenerate and retime the data signals which corrects signal quality loss accumulated through the bypass circuitry between empty slots. Empty backplanes can generate errors which are intermittent and difficult to isolate.
Error message examples:

Jun 6 20:48:11 depth unix: sf0: Target 0x12 Reset Failed. Ret=105
Jun 6 20:48:11 depth unix: sf0: sf_reset all
Jun 6 20:48:11 depth unix: ID[SUNWssa.socal.link.5010] socal0: port 0:
Fibre Channel is OFFLINE
Jun 6 20:48:11 depth unix: WARNING:
/sbus@1f,0/SUNW,socal@0,0/sf@0,0/ssd@w2100002037070d4c,0 (ssd8):
Jun 6 20:48:11 depth unix: SCSI transport failed: reason 'timeout':
retrying command
Jun 6 20:48:11 depth unix: ID[SUNWssa.socal.link.6010] socal0: port 0:
Fibre Channel Loop is ONLINE
Jun 6 20:48:13 depth unix: sf0: ELS 0x0 to target 0x1d retrying
Jun 6 21:00:07 depth unix: WARNING:
/sbus@1f,0/SUNW,socal@0,0/sf@0,0/ssd@w2100002037070d5a,0 (ssd5):
Jun 6 21:00:07 depth unix: Error for Command: read(10) Error Level:
Retryable
I0373-1
n/a
A5000 EMI clip SMCC Engineering has noted several isolated instances involving the Enterprise Network Array, where the EMI clip sandwiched between the underside of the Horizontal Interconnect Board and the top edge of the Vertical Interconnect Board nearest the Interface Boards, had shifted just far enough to allow the EMI clip to overlap and make contact with the soldermask over the Fibre Channel signal traces. If one trace is grounded by the clip, the unit may experience a hard If one trace is grounded by the clip, the unit may experience a hard clip, Loop "A" will experience a hard fault.
This problem can manifest as the following errors on the HOST system:

WARNING: /sbus@2,0/SUNW,socal@d,10000/sf@0,0/ssd@w210000203707d2a3,0 (ssdX):
SCSI transport failed: reason 'timeout': retrying command

WARNING: /sbus@2,0/SUNW,socal@d,10000/sf@1,0/ssd@w210000203707d288,0 (ssdX):
SCSI transport failed: reason 'timeout': retrying command

WARNING: /sbus@2,0/SUNW,socal@d,10000/sf@0,0/ssd@w210000203707d1d3,0 (ssdX):
SCSI transport failed: reason 'timeout': retrying command

WARNING: /sbus@2,0/SUNW,socal@d,10000/sf@1,0/ssd@w210000203707d3e2,0 (ssdX):
SCSI transport failed: reason 'timeout': retrying command


WebToneWebToneWebToneWebTone
 Copyright 1994-2003 Sun Microsystems, Inc.,  901 San Antonio Road, Palo Alto, CA 94303 USA.  All rights reserved.
 Legal Terms Privacy Policy Feedback