A limited number of Sun Fire T2000 and SPARC Enterprise T2000 servers may experience a shutdown with SC Alert: "Chassis cover removed".

Asset ID:	1-73-1020218.1
Update Date:	2010-09-23
Keywords:

Solution Type FAB (standard) Sure

Solution 1020218.1 : A limited number of Sun Fire T2000 and SPARC Enterprise T2000 servers may experience a shutdown with SC Alert: "Chassis cover removed".

Impact

A limited number of Sun Fire T2000 and SPARC Enterprise T2000 servers may experience a system shutdown after the System Controller (SC) Alert: "Chassis cover removed" is displayed on the console, causing system downtime.

Contributing Factors

This issue can occur on the following platforms:

- Sun Fire T2000 Server
- Sun SPARC Enterprise T2000 Server

Note: This issue rarely occurs, and has only been observed on the above mentioned
T2000 servers. No other Sun systems are affected by this issue.

Symptoms

The system will report the following errors on the system console, which will also be recorded in the ALOM logs. An example from 'showlogs -v' would be similar to the following:

NOV 09 02:24:25: 0004007c: "System poweron is disabled."
NOV 09 02:24:25: 00040083: "Chassis cover removed."
NOV 09 02:24:25: 0004000e: "SC Request to Power Off Host Immediately." <<<<<<<<
NOV 09 02:24:26: 0004004f: "Indicator SYS/ACT is now STANDBY BLINK"
NOV 09 02:24:27: 0004007d: "System poweron is enabled."
NOV 09 02:24:31: 00040029: "Host system has shut down."

As shown in the example, the key to identify this issue is that in the logs, the line "Chassis cover removed" will be followed by the line "SC Request to Power Off Host Immediately". If the line "SC Request to Power Off Host Immediately" is missing from the above message, then this is a different issue and may indicate a hardware condition with the cover interlock switch.

Root Cause

The suspected root cause is invalid CI (Chassis Intrusion) bit read from the ADM1026, either caused by i2c corruption or low ADM1026 CI pin noise tolerance. Also, the ALOM shutdown (based on SystemPowerON check) after failed Read from ADM1026 should be disabled, because in a real CI, the FPGA will have already turned off power.

So the poweron check, in conjuction with the root cause (i2c corruption or over-sensitive adm1026 CI pin), causes the host to power off with the message "SC Request to Power Off Host Immediately".

A firmware patch has been developed to permit up to three retry reads to ADM1026, with clear in between to confirm status. If ALOM is still reporting a chassis cover problem after 3 tries, it will display a message, but will NOT shutdown the box.

Corrective Action

Workaround:

On occurrence of the "Chassis cover removed" error, perform a full AC powercycle of the server. Poweroff the server, remove the AC power cords, wait approximately 30 seconds, then plug back the AC power cords and power-on the server. This will reset the I2C bus and clear the error status.

Resolution:

Install patch 139434-02.

References:

BugID: 6780678, 6815610
Escalation ID: 1-25151443, 1-25151473, 1-25325594

For information about FAB documents, its release processes, implementation strategies and billing information, go to the following URL:

http://tns.central/fab

For Sun Authorized Service Providers go to:

http://tns.central/SPE/fab/

In addition to the above you may email:

[email protected]

Modification History
Changes made since initial publication.

06-Apr-2009

Changed step 2 in Corrective Action section from replace Front I/O Board to use listed IDR Patch.

15-Apr-2009

Changed from Preliminary to Resolved, removed step 2 in workaround and added patch id in Resolution.

Internal Contributor/submitter
[email protected]

Internal Eng Responsible Engineer
[email protected] Responsible Manager: [email protected]

Internal Services Knowledge Engineer
[email protected]

Internal Eng Business Unit Group
SSG WGS (Workgroup Systems)

Internal Sun Alert & FAB Admin Info
09-Mar-2009: Completed draft and sent to Extended Review.
11-Mar-2009: Addressed all feedback from Ext Rvw - sending to Publish.

Attachments

This solution has no attachment