Understanding The Error Codes - IBM Storwize V7000 Unified Problem Determination Manual

Understanding the error codes

Error codes are generated by the event-log analysis and system configuration code.
Error codes help you to identify the cause of a problem, a failing component, and
the service actions that might be needed to solve the problem.
Viewing logs and traces
The Storwize V7000 Unified clustered system maintains log files and trace files that
can be used to manage your system and diagnose problems.
You can view information about collecting log files or you can view examples of a
configuration dump, error log, or featurization log. To do this, click Reference in
the left pane of the Storwize V7000 Unified Information Center and then expand
the Logs and traces section.
Understanding the Storwize V7000 Unified battery operation for the
control enclosure
Storwize V7000 Unified node canisters cache volume data and hold state
information in volatile memory.
If the power fails, the cache and state data is written to a local solid-state drive
(SSD) in the canister. The batteries within the control enclosure provide the power
to write the cache and state data to a local drive.
Note: Storwize V7000 Unified expansion canisters do not cache volume data or
store state information in volatile memory. They, therefore, do not require battery
power. If ac power to both power supplies in an expansion enclosure fails, the
enclosure powers off. When ac power is restored to at least one of the power
supplies, the controller restarts without operator intervention.
There are two power supply units in the control enclosure. Each one contains an
integrated battery. Both power supply units and batteries provide power to both
control canisters. Each battery has a sufficient charge to power both node canisters
for the duration of saving critical data to the local drive. In a fully redundant
system with two batteries and two canisters, there is enough charge in the batteries
to support saving critical data from both canisters to a local drive twice. In a
system with a failed battery, there is enough charge in the remaining battery to
support saving critical data from both canisters to a local drive once.
If the ac power to a control enclosure is lost, the canisters do not start saving
critical data to a local drive until approximately 10 seconds after the loss of ac
power is first detected. If the power is restored within this period, the system
continues to operate. This loss in power is called a brown out. As soon as the
saving of the critical data starts, the system stops handling I/O requests from the
host applications, and Metro Mirror and Global Mirror relationships go offline. The
system powers off when the saving of the critical data completes.
If both node canisters shut down without writing the cache and state data to the
local drive, the system is unable to restart without an extended service action. The
system configuration must be restored. If any cache write data is lost, volumes
must be restored from a backup. It is, therefore, important not to remove the
canisters or the power supply units from the control enclosures unless directed to
182
Storwize V7000 Unified: Problem Determination Guide 2073-720