StorageTek Virtual Storage Manager System 6 (VSM6)
 
StorageTek Virtual Storage Manager System 7 (VSM7)
 
Sun Virtual Library Extension (VLE)
 

PLA-Support>Sun Systems>TAPE>Virtual Tape>SN-TP: VSM6
 

In this Document

Symptoms

Changes

Cause

Solution

References

Oracle Confidential PARTNER - Available to partners (SUN).
Reason: Partners and Oracle only
Created from <SR 3-17344314401>

Applies to:

Symptoms

Command timeout errors are being reported

SSD fenced

SSD or JBOD disk status is offline or removed

Common questions related to an SSD or disk failure

Changes

 None

Cause

Faulty JBOD (just a bunch of disks) disk

Faulty SSD (solid state device)

Too many command timeouts for an SSD or JBOD drive (aka disk)

Solution

Command timeouts can occur on an SSD or JBOD disk device for a VSM6, VSM7, or VLE.  The command timeouts are related to a failure of reading or writing to one of these devices.  As with the VLE, the VSM6 and VSM7 subsystems have a monitoring script designed to monitor the command timeouts and offline the SSD or drive.  Should the command timeouts reach the threshold of 5 timeouts within 15 minutes, the SSD or JBOD disk will typically be placed in an offline status.  There are instances where the SSD or disk reports "No offline done. Not able to get the disk status." and the device will not be placed offline but placed in a status of "removed" by the system.  In either of these circumstances, an event in the alarms.txt log will be generated and an ASR created (if configured) identifying the failed SSD or disk.  The VSM6, or VSM7 Installation and Configuration Service Guide document has instructions to replace the failed SSD or disk.  In the case of a an offlined or removed SSD, the SSD will be fenced and there are procedures in this document for this scenario.

 Some common questions related to an SSD or disk failure.

Questions for command timeouts with regards to the SSDs on a VSM6 or VSM7?

1. What is the lifetime expectancy of an SSD?  Answer - There is no known, specific life expectancy for an SSD at the time of the writing of this document.  Upon a failure, replace the SSD.
2. How does the VSM6 or VSM7 control when an SSD or disk is to expire or fail?  There is nothing other than the command timeouts threshold set or monitored within the VLE, VSM6 or VSM7.  When there are 5 timeouts within 15 minutes, the disk will be failed with a status of offline or removed.  Devices that do not exceed the threshold of 5 timeouts within 15 minutes do NOT need to be replaced - they are operating within Engineering specified parameters.
3. Is there a "lifetime of access count for an SSD?  Answer - There are no known "lifetime access counts" for an SSD or disk at the time of the writing of this KB document.

Background History

The Oracle VLE engineering team wrote the script to offline a disk for too many command timeouts to avoided the complexity of automatically picking a spare disk and initiating the resilver process. The purpose of offlining a disk is to stop sending I/O to the failing disk and avoid long delays that can further hang the system operation or cause performance issues. Offlining the disk will not induce drive re-silvering by design, therefore a spare will not be used.  The VSM6 and VSM7 engineering team incorporated this monitoring script to perform the same function as the VLE subsystem.

References

This solution has no attachment