R/Evolution 2000 Series Troubleshooting Manual page 46

Problem PHYs can cause a host or controller to continually rescan drives, which
disrupts I/O or causes I/O errors. I/O errors can result in a failed drive, causing a
virtual disk to become critical or causing complete loss of a virtual disk if more than
one fails.
To avoid these problems, problem PHYs are identified and disabled, if necessary,
and status information is transmitted to the controller so that each action can be
reported in the event log. Problem PHY identification and status information is
reported in RAIDar, but disabled PHYs are only reported through event messages.
Some PHY errors can be expected when powering on an enclosure, when removing
or inserting a controller, and when connecting or disconnecting an enclosure. An
incompletely connected or disturbed cable might also generate a PHY error. These
errors are usually not significant enough to disable a PHY, so the fault isolation
firmware analyzes the number of errors and the error rate. If errors for a particular
PHY increase at a slow rate, the PHY is usually not disabled. Instead the errors are
accumulated and reported.
On the other hand, bad cables connecting enclosures, damaged controller
connectors, and other physical damage can cause continual errors, which the fault
isolation firmware can often trace to a single problematic PHY. The fault isolation
firmware recognizes the large number and rapid rate of these errors and disables this
PHY without user intervention. This disabling, sometimes referred to as PHY
fencing, eliminates the I/O errors and enables the system to continue operation
without suffering performance degradation.
Once the firmware has disabled a PHY, the only way to enable the PHY again is to
reset the affected controller or power cycle the enclosure. Before doing so, it may be
necessary to replace a defective cable or FRU.
If a PHY becomes disabled, the event log entry helps to determine which enclosure
or enclosures and which controller (or controllers) are affected.
RAIDar provides an Expander Status page, which contains an Expander Controller
Phy Detail panel. This panel shows information about each PHY in the internal data
paths between the Storage Controller, Expander Controller, drives, and expansion
ports. By reviewing this page you can quickly locate the internal data path that has
a fault.
46 R/Evolution 2000 Series Troubleshooting Guide • May 2008