Sun Fire X4540 Server
 
Solaris Operating System
 
Sun Fire X4500 Server
 

PLA-Support>Sun Systems>x86>Server>SN-x64: AMD-STOR-SERVER
 

To discuss this information further with Oracle experts and industry peers, we encourage you to review, join or start a discussion in the My Oracle Support Community - Sun x86 Systems

Applies to:

Symptoms

There are instances when an Sun Fire[TM] X4500's drive firmware SMART (Self-Monitoring Analysis and Reporting Technology) predictively fails out a disk and reports it to fmadm.

ZFS however can report the drive as healthy.

Running cfgadm -c unconfigure is how the service manual recommends replacing the drive. However, since the drive was still healthy according to ZFS, the command will fail with the following:

For patched Solaris 10, ZFS now recognizes the predictive failure reported via SMART and faults the disk automatically.

The fix is in the Solaris 10 patch for <Bug: 15662359> activate ZPOOL_CONFIG_FRU in ON

Patches:

149637-04 - x86
150125-03 - SPARC

Solaris 11.x has already implemented the feature.

Changes

N/A

Cause

.N/A

Solution

Since the drive was still healthy according to ZFS it needs to be offlined, cfgadm unconfigured, physically replaced, cfgadm configured, and finally zpool replaced

1. Prior to replacing the drive, cfgadm -alv , will show the following output: 

(output ommitted)

 2. fmadm and fmdump will show the drives as faulty: 

3. Format will show the following: 

(output ommitted)

 4. The zfs commands zpool will show that the pool is healthy and online: 

5. In order to replace the drive, you need to offline the drive in zfs:

 6. The zpool status command will show the following after the drive has been offlined.

7. Now that the drive has been offlined from the zfs pool, it can be removed from dynamically reconfigured from OS control by running the following command:

 8. Notice that the drive no longer shows up in cfgadm

 9. At this point the drive is safe to remove. You should see the drive's blue light lit up indicating that it is safe to remove it. Physically Replace Drive.

10. Once the drive has been physically replaced, you can configure it back into OS control by running the following command:

 11. Notice that cfgadm now shows the drive again:

 12. You can now put the drive into zfs control by running the below substituting your drive's c#t#d# for c1t7d0 in the example below.

 13. The pool is now healthy again: 

Notice the message that in the zpool status command below under scrub:

14. Use the fmadm command to repair the status of the drive in the fault management service: 

Previously Published As 88150

This solution has no attachment