Sun StorEdge 3510 in Redundant Controller Configuration May Experience Higher Rate of Controller Failover Under Heavy I/O Load

---

Impact

The Sun StorEdge 3510 Fiber-Channel (FC) Array without array controller firmware version 3.27N or later in a "redundant controller" configuration may experience a higher rate of controller failovers. This may render the array temporarily unavailable.

Contributing Factors

This issue can occur on the following platforms:

SPARC Platform

• Sun StorEdge 3510 FC Array without array controller firmware version 3.27N

Note: Firmware 3.27N is available with patch 113723-04 or later.

This issue is prevalent when "cache mode" is set for Random Optimization, however, Sequential Optimization is also subject to an identical condition. To determine the "cache mode" from the Sun StorEdge 3510 main menu, perform the following steps:

1. Select "view and edit Configuration parameters"
2. Select "Caching Parameters"
3. See if "Optimization for Random I/O" is selected

Or, from the Sun StorEdge Configuration Service Command Line Interface (sccli(1M)), the following would be the result of running the "cache-policy" command:

    # /opt/SUNWsade/sbin/sccli cache-policy
    optimization: random

Symptoms

The controller failure may occur with large block sequential writes on RAID-0 and RAID-1 devices with large capacity LDEVs (Logical Devices), i.e. those containing 10 or more disk drives.

Should the described issue occur, messages similar to the following are logged to the "/var/adm/messages" file:

    modo SUNWscsdMonitor[19670]: [ID 254907 daemon.error] [SUNWscsd 0x10B1D03:
    Critical] <rctrl5040> Controller
    Event, Redundant Controller Failure Detected
    Likely controller error. If error persists, replace defective controller.

Or, from the Sun StorEdge 3510 "Main Menu", messages similar to the following alert may be displayed:

    [0111] Controller ALERT: Redundant Controller Failure Detected

Or, by running the sccli(1M) command "show events", the following message will be displayed:

    # /opt/SUNWsade/sbin/sccli show events
    [0111]#6:StorEdge Array SN#3341163 Controller ALERT:redundant controller failure detected

Or, the GUI will display the error code "010b1d03" (redundant controller failed).

Workaround

There is no workaround. Please see the "Resolution" section below.

Resolution

SPARC Platform

• Sun StorEdge 3510 FC Array with array controller firmware version 3.27N (delivered with patch 113723-04 or later)

Modification History


Date: 29-JUL-2003

• State: Resolved
• Updated Synopsis
• Updated Contributing Factors
• Updated Resolution

Attachments

This solution has no attachment