IBM System Storage DS4000, Руководство по оборудованию

Страница: 423 / 566

Chapter 7. Advanced maintenance, troubleshooting, and diagnostics
405
7.7.8 Pre-read redundancy check
The pre-read redundancy check verifies the redundancy information during every read I/O. A
logical drive that has this feature enabled returns read errors if the data is determined to be
inconsistent by the controller. You can enable this option for logical drives that contain
redundancy information, that is, RAID 1, 3, 5, and 6. However, there is a severe performance
impact with this feature enabled, so it is only enabled in exceptional cases.
7.8 Problem determination
In this section, we explore some of the tools and methods used to diagnose:
򐂰 Drive-side problems
򐂰 Host-side problems
򐂰 Storage Manager communication problems
We only include techniques for determining faults using the Storage Manager and SMCli
interface. Additional tools are available through the controller shell, although these are
outside the scope of this book. A password is required for accessing the shell through either
the serial port or a telnet session. It is only intended to be used by the IBM Support
representative during problem determination.
However, there is a Capture State Information option in Storage Manager (also included
within the Collect all Support Data bundle) that collects shell data from both controllers in the
DS4000 or DS5000 storage subsystem. We will explore some of the included commands that
might be relevant to fault finding.
With any suspected fault, it is essential to first rule out the possibility of a code mismatch or
configuration error. This is particularly important when fault symptoms are detected following
an installation or upgrade. Check that all prerequisites and configuration rules are adhered to
before logging a hardware support call.
In many cases, Recovery Guru will assist in identifying cabling configuration errors with an
explanation of the fault together with advice on correcting it, for example:
Channel miswired Two or more drive channels are connected to the
same Fibre Channel loop. The Recovery Guru
Details area provides specific information that
you will need as you follow the recovery steps.
Drive enclosures not cabled correctly There are drive enclosures in the storage
subsystem that are not cabled correctly because
they have ESM canisters that must be cabled
sequentially together. The Recovery Guru
Details area provides specific information that
you will need as you follow the recovery steps.
For limitations and version specific requirements, see the readme documents associated with
the current code files, which can be found at the following address:
http://www.storage.ibm.com/support
For details of tested and supported configurations, see the System Storage Interoperation
Center at the following address:
http://www-03.ibm.com/systems/support/storage/config/ssic