Access to Data on a Sun StorEdge T3/T3+ and Sun StorEdge 6120 Array May be Temporarily Lost When a Substituted Drive is Replaced

---

Impact

On a Sun StorEdge T3/T3+ or Sun StorEdge 6120 array, access to data may be lost when a drive that has previously failed (and has been logically disabled and substituted) is removed. The data will be inaccessible until the volume is remounted and any necessary recovery actions are performed on the host. Since the loss of access is abrupt, loss of data not yet committed to disk is possible.

Contributing Factors

This issue can occur in the following releases:

SPARC Platform

• Sun StorEdge T3 Array with firmware 1.01a through 1.16c (as delivered with patch 109115-01 through 109115-06)
• Sun StorEdge T3 Array with firmware 1.18.0 through 1.18.4 (as delivered with patch 109115-10 through 109115-17)
• Sun StorEdge T3+ Array without firmware 3.1.0 (as delivered with patch 115180-01)
• Sun StorEdge 6120 Array without firmware 3.1.0 (as delivered with patch 115179-01)

The Sun StorEdge "ver" command at the T3, T3+ or 6120 system prompt can be used to determine the current firmware revision, as shown in the following example:

    6120a:/:<1>ver
    6120 Release 3.1.0 Fri Dec 19 14:49:54 PST 2003 (129.156.169.50)
    Copyright (C) 1997-2003 Sun Microsystems, Inc

Note: In order for the described issue to occur, a drive failure must first result in the successful substitution of the hot spare followed by a second drive failure in the same RAID set. On a RAID 5 volume, this issue can occur if a second drive fails in any slot. On a RAID 1 volume, this issue will only occur if the second drive failure is adjacent to the first failed drive.

Symptoms

The "vol stat" command shows one drive substituted and another drive failed in a volume. When the drive in the substituted state is removed, the volume is unmounted.

Workaround

To work around the described issue, do the following:

1. Replace the second failed drive and keep the substituted drive in place. Do not remove the substituted drive until the reconstruction process is complete.

2. Replace the first failed drive which was substituted by the standby drive. The volume is fixed once this reconstruction process is complete.

Note: Before replacing either failed drive, be sure to backup the data on the affected volume.

Resolution

This issue is addressed in the following releases:

SPARC Platform

• Sun StorEdge T3+ Array with firmware 3.1.0 (as delivered with patch 115180-01) or later
• Sun StorEdge 6120 Array with firmware 3.1.0 (as delivered with patch 115179-01) or later

Note: The "Relief/Workaround" provided above is the final resolution for the Sun StorEdge T3 array.

Modification History




Attachments

This solution has no attachment