Possible NAS panic when attempting to recover a failed port of an HA bond while there is an active iSCSI connection to a LUN on the NAS.

Asset ID:	1-73-1000929.1
Update Date:	2010-11-03
Keywords:

Solution Type FAB (standard) Sure

Solution 1000929.1 : Possible NAS panic when attempting to recover a failed port of an HA bond while there is an active iSCSI connection to a LUN on the NAS.

Related Items


Sun Storage 5220 NAS Appliance
 Sun Storage 5210 NAS Appliance
 Sun Storage 5310 NAS Appliance
 Sun Storage 5320 NAS Appliance

Related Categories


GCS>Sun Microsystems>Sun FAB>Standard>Reactive

PreviouslyPublishedAs
201245

Bug Id
<SUNBUG: 6607042>

Date of Resolved Release
07-Aug-2008

Product
Sun StorageTek 5320 NAS Appliance
Sun StorageTek 5310 NAS Appliance
Sun StorageTek 5220 NAS Appliance
Sun StorageTek 5210 NAS Appliance

Microsoft 2.05 iSCSI initiator causes panic and reboot (see details below).

Impact

The NAS head will panic and reboot.

The purpose of this FAB is to notify the field of the potential issue regarding the use of Microsoft's iSCSI initiator over bonded links.

Contributing Factors

This issue has only been seen with the Microsoft 2.05 iSCSI initiator during connection re-establishment.

Symptoms

The NAS head will panic and reboot. The dbglog on the machine should have a stack trace that looks similar to the following:

  eip 07BD3201  eflags 00010256 ..R...I..ZAP.  vec 0E Page Fault
eax 00000000  ebx 07B40018  ecx 003C97B6  edx 0FCB2440
edi 0852AC90  esi 07B40000  ebp 00FBDE68  esp 00FBDE10
7BD3201 movb   0x6C(%eax),%al     8A406C       iscsiTarget.wscsi_do_inquiry+505

STACK TRACE:

  0xFBDE68 iscsiTarget.wscsi_do_inquiry+505(FCB2440 852AC90 7BF2C2C 7BC3719)
0xFBDE98 iscsiTarget.wscsi_do_inquiry+269(FCB2440 852AC90 7BF2C2C 7BB2FC6)
0xFBDF08 iscsiTarget.wscsi_initial_packet+71(FCB2440 852AC90 10000 7BF2C2C)
0xFBDF88 iscsiTarget.iscsi_ack_StatSN+45C9(7BF2C2C 7B751E0 FCB2440 852AC90)
0xFBDFC8 iscsiTarget.target_worker_t+2CD(7BF2C2C DF5F14 80 769F000)
0xFBDFF8 thread_launcher+0x34(FC6000 0 1155A0 FC5F0C)

Root Cause

A SCSI inquiry command received from the Microsoft iSCSI initiator does not contain a LUN number which matches any of the LUN numbers currently stored on the NAS. This causes an indexing error which leads to a page fault.

The code has been modified to prevent the indexing error by defaulting to LUN number zero (0) when a SCSI inquiry command is received without a valid LUN number.

The issue will be resolved with the release of StorageTek NAS OS version 4.22 M0.

Workaround

If using Microsoft's iSCSI initiator v2.05, do not establish iSCSI connections to the NAS over bonded links.

Resolution

Upgrade NAS OS to release 4.22 M0, which can be acquired via the below links;

StorageTek 5320 NAS software update:

https://sspatch.oracle.com/readme.do?assetkey=1-21-119352-10-1

StorEdge 5310 NAS software update:

https://sspatch.oracle.com/readme.do?assetkey=1-21-119351-15-1

StorageTek 5220 NAS software update:

https://sspatch.oracle.com/readme.do?assetkey=1-21-119353-07-1

StorEdge 5210 NAS software update:

https://sspatch.oracle.com/readme.do?assetkey=1-21-118216-15-1

Reference Sun StorageTek NAS OS Administration Guide (819-4284-11), page 191: Upgrading NAS Software

Identification of Affected Parts (how to):

This issue is found in all releases of the NAS OS from 4.12 M0 to 4.21 M2 including:

   StorageTek NAS OS Version 4.12 M0
StorageTek NAS OS Version 4.20 M0
StorageTek NAS OS Version 4.20 M1
StorageTek NAS OS Version 4.20 M2
StorageTek NAS OS Version 4.20 M3
StorageTek NAS OS Version 4.21 M0
StorageTek NAS OS Version 4.21 M1
StorageTek NAS OS Version 4.21 M2

The NAS OS version can be found on the main panel of the Web Admin or by issuing the 'version' command from the console.

Previously Published As
103146
Internal Contributor/submitter
William.Krier@Sun.COM

Internal Eng Responsible Engineer
Stephen.Ostrowski@Sun.COM Responsible Manager: Martie.Land@Sun.COM

Internal Services Knowledge Engineer
Joe.Davis@Sun.COM

Internal Eng Business Unit Group
NWS (Network Storage)

Internal Kasp FAB Legacy ID
103146

Internal Sun Alert & FAB Admin Info
Significant Change Date: 2007-11-12

06-Nov-2007: Rcvd submission but requested some clarifications from the submitter.
08-Nov-2007: Submitter provided clarifications, draft started and sent out for Ext Rvw.
12-Nov-2007: no feedback from Ext Rvw, sending to Pub.
07-Aug-2008: Changed from Workaround to Resolved.

Attachments

This solution has no attachment