Sun Microsystems, Inc.  Oracle System Handbook - ISO 7.0 May 2018 Internal/Partner Edition
   Home | Current Systems | Former STK Products | EOL Systems | Components | General Info | Search | Feedback

Asset ID: 1-72-1614069.1
Update Date:2014-11-26
Keywords:

Solution Type  Problem Resolution Sure

Solution  1614069.1 :   Sun Storage 2500, 2500-M2, and 6000 Arrays: Solaris Explorer Data Collector Can Trigger write/read Retryable Error On System using Symantec VxDMP  


Related Items
  • Explorer Data Collector
  •  
  • Solaris Operating System
  •  
  • Sun Storage 2540-M2 Array
  •  
  • Sun Storage 2540 Array
  •  
  • Sun Storage 6130 Array
  •  
  • Sun Storage 2530-M2 Array
  •  
  • Sun Storage 2530 Array
  •  
  • Sun Storage 6140 Array
  •  
Related Categories
  • PLA-Support>Sun Systems>DISK>Storage Drivers>SN-DK: Storage Drivers
  •  




In this Document
Symptoms
Changes
Cause
Solution
References


Created from <SR 3-8249029271>

Applies to:

Sun Storage 6130 Array - Version All Versions and later
Solaris SPARC Operating System - Version 10 10/09 U8 and later
Explorer Data Collector - Version 5.11 to 6.4 [Release 5.0 to 6.0]
Sun Storage 2540-M2 Array - Version Not Applicable and later
Sun Storage 2530-M2 Array - Version Not Applicable and later
Oracle Solaris on x86 (32-bit)
Oracle Solaris on x86-64 (64-bit)
Oracle Solaris on SPARC (64-bit)
Oracle Solaris on SPARC (32-bit)

Symptoms

Running Solaris Explorer Data Collector can trigger following type of error message on Solaris system using Symantec VxDMP instead of Solaris Multipathing (MPxIO):

Dec 26 13:56:32 paris scsi: [ID 107833 kern.warning] WARNING: /pci@1d,700000/SUNW,qlc@1,1/fp@0,0/ssd@w200700a0b8170e2d,0 (ssd3):
Dec 26 13:56:32 paris        Error for Command: read(10)                Error Level: Retryable
Dec 26 13:56:32 paris scsi: [ID 107833 kern.notice]         Requested Block: 350909                    Error Block: 350909
Dec 26 13:56:32 paris scsi: [ID 107833 kern.notice]         Vendor: SUN                                Serial Number:    +  $%R:
Dec 26 13:56:32 paris scsi: [ID 107833 kern.notice]         Sense Key: Unit_Attention
Dec 26 13:56:32 paris scsi: [ID 107833 kern.notice]         ASC: 0x8b (<vendor unique code 0x8b>), ASCQ: 0x2, FRU: 0x0

Please refer to Doc 1006536.1 for a detailed explanation of the error.

Changes

 The problem always existed and is reproducible on following configuration:

  • Sun Storage array 6100 and 2500 series
  • Solaris 10u8 and later
  • Solaris Multipathing (MPxIO) disabled
  • Symantec VxVM/DMP 5.x and below or 5.1 and later without option dmp_native_multipathing enabled
  • Symantec VxDMP using libvxengenio.so library for the STK array

Cause

These errors are seen only under the following conditions:

  • the logical unit is accessed via at least one path on each controller
  • AVT occured previously and all path are accessible and respective Solaris disk driver instance (sd/ssd) configured (meaning no path will be shown "Drive Type Unknown" in format disk list)
  • there's a lot of iops on the "active" path at the time Solaris Explorer is taken

Of all the command ran by Solaris Explorer only fstyp and format (even with NOINUSE_CHECK set to 1) command will generate the error.

What is happening is that the command will try to read the label on the "standby" path while i/o are still being made on the active path by the dd command.
On Sun Storage array 6100 and 2500 series, when the host mode is set to "Solaris MPxIO" or "Solaris with DMP", access to read the label which is stored at LBA 0 of the disk does not trigger failover. However since the introduction of Extended VTOC more data is read and this cause a failover to occur.

Due to lack of time it was not clearly seen if this problem was reproducible before S10u8 and introduction of Extended VTOC and fewer changes for EFI support in sd driver.
What was clearly showned at customer site and in lab is that the problem is not reproducible using prtvtoc since it open the sd device with different flag. Only fstyp/format will trigger some specific ioctl that will generate the AVT part of the open() call used to open the sd device.

Solution

This type of error generated when using system command on the standby path will never be fixed at the Solaris OS level as all disk administrative command are clueless about disk storage path state.
Therefore the solution are:

  • enable Solaris Multipathing (see Doc 1511153.1)
  • install Symantec SFS/VxVM 5.1 and later and enable dmp_native_multipathing. Please refer to Symantec documentation for detailed procedure.

References

<NOTE:1006536.1> - Read and Write Operations to a Sun Storage 6000 or 2540 Array Error with ASC 0x8b / ASCQ 0x2
<BUG:15545322> - SUNBT6811234 "FORMAT" COMMAND CAN CAUSE UNWANTED FAILOVER OF ST6XXX/ST25XX VOLUM
<BUG:15509098> - SUNBT6749586 UNDESIRED FAILOVERS DUE TO EFI LABELS.
<NOTE:1511153.1> - SAN - Introduction to MPxIO (Oracle Storage Multi-Pathing Software, AKA STMS, SCSI_VHCI ) and Troubleshooting Central Link Page

Attachments
This solution has no attachment
  Copyright © 2018 Oracle, Inc.  All rights reserved.
 Feedback