Sun Storage 6920 System
 

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

Hosts utilizing Sun StorEdge 6920 arrays may lose access to the array during a volume snapshot device failure.

This issue can occur on the following platforms:

• Sun StorEdge 6920 without Data Services Platform (DSP) 1000 firmware version 3.0.1.13

Should the described issue occur, StorADE for the Sun StorEdge 6920 array will report messages similar to the following:

    LMay 18 10:24:55 dsp00  05/18/2005 09:24:47 LOG_CRIT     (SNAPSHOT: 0-0)
    Snapshot (vol/SRVol-02) is 100 percent full.
    LMay 18 10:30:11 dsp00  05/18/2005 09:30:02 LOG_CRIT     (VSM: 4-4)
    Snapshot failure (snpIOIntentInstWrite/snpIOContextCopyOpWrite), WWN =
    fffffff0-0002c400-00002c40-00000706 (dev = 0x821

The error messages above may be followed by application failure messages.

To work around the described issue, increase the percentage change that is allowed for the snapshot by using either the "Configuration Services Graphical User Interface", or from the sscs(1M) command line. This process requires that the existing snapshot, and snapshot reserve space be deleted.

Note:  The below workaround will not address snapshot failures that are the result of a hardware component failure of the Snapshot Reserve Pool.

For Configuration Services:

1. Input your Service Processor URL into a web browser.

2. Log into the Service Processor GUI.

3. Click on the "Configuration Services" link.

4. Click on the "Logical" Tab.

5. Click on the "Volumes" Tab.

6. Click on the source volume name for the snapshot(s) to get to the "Volume Details" page for this volume.

7. Scroll Down to the "Additional Information" section and click on the "Snapshots" link.

8. Click on each snapshot to access the volume details for that snapshot, and then do the following:

a. Click "Unmap" if highlighted

b. Click "Delete"

9. Go back to the "Volume Details" page for the source volume, and click the "Delete Snapshot Reserve" button.

10. Click on the "Snapshot" button to open the "New Snapshot Creation Wizard".

11. For Step 1, enter the new snapshot name and description in the appropriate fields and click "Next".

12. For Step 2, enter the number of snapshots, and set the % difference to 100 and click "Next".

Note: As indicated, a level(-L) of "full" allocates storage space equal to the source volume specified. The amount of space is calculated as:

    Snapshot Count * Size of Source = Snapshot Reserve Space

13. If you wish to map the snapshot to an initiator, do so in Step 3 and click "Next".

14. Review and click "Finish".

This will create a single snapshot. To create another snapshot using the GUI, click on the "Snapshot" button, and repeat the above steps.

Note: The "Configuration Services" help menu provides additional information about the snapshot utility.

    Help >> Configuring System >> Configuring Volume Snapshot >> Managing Volume Snapshots

From the sscs(1M) command line:

1. login to sscs(1M):

    % sscs login -h <management host> -u storage

2. Unmap the snapshot volume:

    % sscs unmap -S <domain> volume <snapshot volume>

3. Delete the snapshot volume:

    % sscs delete -S <domain> volume <snapshot volume>

4. Delete the snapshot reserve:

    % sscs snapshot -r volume <source_volume>

5. Create new snapshot reserve with 100% space:

    % sscs snapshot -v <source_volume> -C <#snapshots> -L full volume <snapshot volume name>

This will create a single snapshot with the above name, for additional snapshots, do the following:

    % sscs snapshot -v <source volume> volume <new_snapshot_name>

Note: As indicated, a level(-L) of "full" allocates storage space equal to the source volume specified. The amount of space is calculated as:

    Snapshot Count * Size of Source = Snapshot Reserve Space

Use the man page for sscs(1M) to supplement this procedure, if required.

This issue is addressed in the following releases:

• Sun StorEdge 6920 with Data Services Platform (DSP) 1000 firmware version 3.0.1.13 or later

• State: Resolved
• Updated Contributing Factors and Relief/Workaround sections

This issue occurs because the ASC/ASCQ code returned for the snapshot failure causes the host pathing software to retry and fail in a shorter period of time than that which is required for the DSP to reconfigure the ports after the snapshot failure.

We will now return a different, more generic code which will cause the host to timeout and retry at a much longer interval (up to ten minutes).

This solution has no attachment