MediaStore5100 and 5000 series troubleshooting and maintenance
• Persistent Login Failures: A warning alarm (yellow) is generated for any drive that experiences
persistent Login failures. If a Spectrum server cannot log in to a disk for more than five seconds, that
drive will be temporarily removed from the RAID set (subject to the conditions mentioned above), but
no Spectrum server in the RAID set will attempt to start a rebuild on the drive's RAID set, even if a hot
spare is available. If the login failure persists for a period of two minutes, or if any write operation to
the disk in question is postponed for at least 30 seconds, then the disk will be failed, regardless of any
of the requisite conditions mentioned above. Note that in extreme circumstances, this may cause the
Spectrum server to shut down its file system, stopping all playback and recording on that Spectrum
server. If the drive successfully logs in before these limits are reached, an automatic “surgical” rebuild
will be immediately started. Otherwise, one of the Spectrum servers in the system will automatically
start a rebuild (provided a hot spare is available.)
• Write Errors: Warning alarms (yellow) are generated for any drive that reports write errors. Write errors
are errors that can occur occasionally on any system and do not, by themselves, imply catastrophic
drive failures. Drives exhibiting 32 read/write errors (both hard and soft errors) will attempt to be auto-
failed by the Spectrum server provided the RAID Set criteria mentioned above is met. Refer to "About
bad-block auto-repair" for more information.
• Read Errors: Warning alarms (yellow) are generated for any drive that reports read errors. Read errors
are errors that can occur occasionally on any system and do not, by themselves, imply catastrophic
drive failures. Drives that exhibit any combination of read/write errors at different disk drive addresses
will attempt to be auto-failed by the Spectrum servers provided the RAID Set criteria mentioned above
is met. Refer to "About bad-block auto-repair" for more information.
• Auto-Fail Starting: Failure alarms (orange) are generated for any drive that exhibits a problem
requiring it to be removed from the RAID Set. This alarm will report that the Auto-fail process is
queued and provide the reason for the attempted auto-failure.
• Auto-Fail Succeeded: Warning alarms (yellow) are generated for any drive that is successfully
removed from the RAID Set following the auto-fail starting alarm. If this alarm is generated, the drive
will already have been failed. A rebuild should be started on the RAID Set, and this drive should be
removed from the MediaStore and replaced with a good drive.
• Auto-Fail Not Successful: Critical alarms (red) are generated in the for any drive that is not
successfully removed from the RAID Set following the auto-fail starting alarm. If this alarm is
generated, a reason is given and user intervention is needed to correct the situation. Drive auto-fail
attempts are retried in five-minute intervals until the drive is successfully failed or the error condition
has been resolved. Reason codes and suggested user actions are as follows:
• RAID Set Compromised: A hot spare needs to be added to the RAID Set to allow a rebuild of the RAID
Set. Once the rebuild has completed, an additional hot spare should be added to the RAID Set to allow
for the failure of the bad drive.
• RAID Set Rebuilding: A hot spare should be added to the RAID Set to allow the bad drive to be auto-
failed once the rebuild has completed.
• Bad Blocks in RAID Set: Bad blocks on other drives in the RAID set can cause the rebuild to stall,
which could result in data loss. As a result, a drive cannot be auto-failed if there are other drives in the
RAID set with bad blocks. Once the bad blocks are auto-repaired, the drive should be auto-failed by
the Spectrum server. Refer to "About bad-block auto-repair" for more information.
• Disk Readiness Test Failure: A warning alarm (yellow) is generated for any drive that fails a Disk
Readiness Test. Every four hours approximately, each disk in the system is tested for readiness.
This test is transparent for disks that are currently in use, but allows any disk that is otherwise idle to
indicate whether it believes itself ready for use. A drive may respond to this by generating a SMART
warning or Hardware error, which are discussed above, but it may also fail to respond at all. In this
case, the Spectrum servers will report a Disk Readiness Test Failure.
• Insubordinate Disk: A warning alarm (yellow) is generated for any drive that becomes insubordinate.
In this rare case, a disk refuses to accept commands from any Spectrum server until it is power-
213
Spectrum System 8.4 Installation Guide
