Sun Storage 6130 Array
 

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

Unmapping a volume or snapshot on a Sun StorEdge 6130 Array may cause one of the array controllers to reset.

The reset of the Sun StorEdge 6130 array controller will cause all volumes owned by that controller to failover to the remaining controller. Provided multipathing software is correctly configured on the attached hosts, any I/O to those volumes will resume after failover. After the controller reset has completed (usually in less than 1 minute) the volumes can be moved back to their preferred controller.

This issue can occur on the following platform:

• Sun StorEdge 6130 Array with controller firmware 06.12.09.10 and without controller firmware 06.12.10.11 (as delivered in patches 118185-15 and 117856-19)

Note: Controller firmware release 06.12.09.10 is supplied with v1.3 of the Sun StorEdge 6130 Array Management Host Software.

If the described issue occurs, the array's log file (Major Event Log) will show a controller reset occurring at the same time as the "unmap" of the volume or snapshot as shown in the following example:

    Date/Time: Thu Jan 19 07:49:18 MET 2006
    Sequence number: 287858
    Event type: 5210
    Event category: Command
    Priority: Informational
    Description: Delete volume-to-LUN mapping
    Event specific codes: 0/0/0
    Component type: Volume
    Component location: TESTVOLUME
    Logged by: Controller in slot B
    Raw data:
    .....
    Date/Time: Thu Jan 19 07:49:18 MET 2006
    Sequence number: 287859
    Event type: 6104
    Event category: Notification
    Priority: Informational
    Description: This controller's alternate was failed
    Event specific codes: 0/0/0
    Component type: Controller
    Component location: Controller in slot B
    Logged by: Controller in slot B
    Raw data:
    .....
    Date/Time: Thu Jan 19 07:49:18 MET 2006
    Sequence number: 287860
    Event type: 400F
    Event category: Notification
    Priority: Informational
    Description: Controller reset by its alternate
    Event specific codes: 0/0/0
    Component type: Controller
    Component location: Controller in slot A
    Logged by: Controller in slot B
    Raw data:
    .....

Note: The Major Event Log is gathered by the 6130SupportData utility.

To work around the described issue, ensure that all volumes and snapshots on the array are mapped to either hosts or hostgroups. Do not leave any volumes or snapshots unmapped.

To resynchronise information for volumes or snapshots that have already become unsynchronised, map all volumes and snapshots to either hosts or hostgroups and reboot the array.

This issue is addressed in the following release:

• Sun StorEdge 6130 Array with firmware 06.12.10.11 or later

Controller firmware release 06.12.10.11 is supplied by the following patches:

• 118185-15 6130 services Release (6130 Upgrader/Downgrader)
• 117856-19 6130 Baseline Firmware Release (FC CSM Array Controllers)

Note: Both of the above patches have special install instructions included in the README files with each patch.

• State: Resolved
• Updated Contributing Factors and Resolution Sections

• Updated Contributing Factors and Resolution sections

References

The bugID listed in the Sun Alert is not listed in the patch README as being addressed.  Sent e-mail below to the submitter to get the patch README updated.

Hi Brian,

The Sun Alert team are complaining that the patch README file for 117856-19 does not mention that it fixes bug 6369690 in the "BugId's fixed with this patch:" section of the file. Can you get the patch README file updated please?

Thanks.

-------- Original Message --------

Subject: Re: Update request for UNRESOLVED Sun Alert 102191

Date: Wed, 08 Mar 2006 11:16:44 -0700

From: Jeff Folla

To: [email protected]

Hi Matt,

The bugID listed in the Sun Alert is not listed in the patch READMEs for patch 117856-19 as being addressed.

Can you get the patch README updated to include bugID 6369690 as being fixed? Once this is done I can update the Sun Alert as resolved with the installation of these patches.

------------

The information that can become unsynchronised for the volume/snapshot is the "LUNMAPG_REF". This can be viewed in the output of "spmShow". The 6130SupportData utility gathers the output of "spmShow" from both array controllers in the "stateCaptureData.dmp" file. Look for the two outputs headed:

     Controller in slot B

     Executing spmShow(0,0,0,0,0,0,0,0,0,0):

 and

     Controller in slot A

     Executing spmShow(0,0,0,0,0,0,0,0,0,0):

 

Compare the information in the last section of these two outputs (headed "LUM") to see if the "LUNMAPG_REF" values are synchronised between the two controllers.

Here is an example :

Controller in slot A

     Executing spmShow(0,0,0,0,0,0,0,0,0,0):

     ...

     LUM

     LUNMAPG_REF LUN SSID VOLUME_REF/(GV - WWN) PERMS

     MAP_REF MAP_REF_DESCRIPTION

     880104000000 0 1024 0000000000000000000000000000000000000000 0xf

     840101000001 Host =B-Controller

     88030D000002 0 0 02000000000017D643D9CEC70000000000000000 0x0

     840301000002 Host =my-host

     880105000006 0 1024 0000000000000000000000000000000000000000 0xf

     840101000000 Host =A-Controller

     880105000007 1 1 020000000000164743D9CEC20000000000000000 0x0

     000000000000 Host =my-host

     880104000008 2 2 02000000000017D843D9CF570000000000000000 0x0

     000000000000 Host =my-host

 

Controller in slot B

     Executing spmShow(0,0,0,0,0,0,0,0,0,0):

     ...

     LUM

     LUNMAPG_REF LUN SSID VOLUME_REF/(GV - WWN) PERMS

     MAP_REF MAP_REF_DESCRIPTION

     880104000000 0 1024 0000000000000000000000000000000000000000 0xf

     840101000001 Host =B-Controller

     880129000001 1 1 020000000000164743D9CEC20000000000000000 0x0

     000000000000 Host =my-host

     88030D000002 0 0 02000000000017D643D9CEC70000000000000000 0x0

     840301000002 Host =my-host

     88011B000003 2 2 02000000000017D843D9CF570000000000000000 0x0

     000000000000 Host =my-host

     880105000006 0 1024 0000000000000000000000000000000000000000 0xf

     840101000000 Host =A-Controller

 

Here the "LUNMAPG_REF" for the volume with SSID 0 is the same on both controllers - "88030D000002" - but the "LUNMAPG_REF" for the other two volumes has become unsynchronised between the two controllers:

     Volume with SSID 1 - On A = "880105000007", On B = "880129000001"

     Volume with SSID 2 - On A = "880104000008", On B = "88011B000003"

 

If either of these two volumes were unmapped, then a controller reset would occur.

PTS Reviewer (approved by): Sailesh Thanki

References

This solution has no attachment