Sun SPARC Enterprise T5440 Server
 
SPARC T3-1B
 
SPARC M5-32
 
SPARC T3-4
 
SPARC T5-1B
 
Netra SPARC T4-1 Server
 
SPARC T4-2
 
Sun Netra T5220 Server
 
Sun SPARC Enterprise T5220 Server
 
Sun SPARC Enterprise T5240 Server
 
Sun Blade T6320 Server Module
 
Sun Netra T5440 Server
 
Netra T3-1
 
Netra SPARC T5-1B Server Module
 
Sun Blade T6340 Server Module
 
SPARC T4-1B
 
Sun Netra T6340 Server Module
 
SPARC T5-2
 
SPARC T3-2
 
SPARC T4-1
 
SPARC T3-1
 
Sun SPARC Enterprise T5140 Server
 
Netra SPARC T4-2 Server
 
SPARC T5-4
 
Sun SPARC Enterprise T5120 Server
 
Netra SPARC T4-1B
 
SPARC T4-4
 
SPARC T5-8
 

PLA-Support>Sun Systems>Sun_Other>Sun Collections>SN-OTH: Sun Alert
 
_Old GCS Categories>Sun Microsystems>Sun Alert>Release Phase>Resolved
 

In this Document

Description

Occurrence

Symptoms

Workaround

Patches

History

References

Applies to:

Description

An issue with SPARC T-series and M5-32 Sun System Firmware may cause OS memory corruption on a host running Solaris 10 with patch 142909-17 or Solaris 11. This memory corruption may lead to data loss and/or unexplained system or application crash.

Occurrence

This issue can occur on the following platforms:

The following 2 key factors must BOTH  be present for the issue to occur:

    1. System is running Solaris 10 with patch 142909-17 or Solaris 11

AND:

    2. The system firmware revision level is earlier than shown below.

To determine if the Solaris 10 patch 142909-17 is installed, use the following command:

      $ showrev -p | grep "Patch: 142909-17"

Note that systems installed with Solaris 10 9/10 or 8/11 or 1/13 will all show that 142909-17 is installed.  Earlier releases of Solaris 10 may or may not have this patch present on the system.

Use the ILOM or Solaris prtdiag(1M) to determine firmware revision level (2):

Log into the Service Processor and run:

      -> show /HOST sysfw_version

         /HOST  
         Properties:
           sysfw_version = Sun System Firmware 9.0.0.i 2013/05/16 17:45

Or from Solaris:

      # prtdiag -v | grep Firmware
      Sun System Firmware 9.0.0.i 2013/05/16 17:45

      SPARC M5-32                        9.0.2.e
      SPARC T5-8                           9.0.0.k
      SPARC T5-4                           9.0.0.k
      SPARC T5-2                           9.0.0.k
      SPARC T5-1B                         9.0.0.k
      Netra SPARC T5-1B                9.0.0.k
      Netra SPARC T4-2 Server      8.3.0.d
      Netra SPARC T4-1B               8.3.0.d
      Netra SPARC T4-1 Server      8.3.0.d
      SPARC T4-4                          8.3.0.d
      SPARC T4-2                          8.3.0.d
      SPARC T4-1B                        8.3.0.d
      SPARC T4-1                          8.3.0.d
      Netra SPARC T3-1 Server     8.3.0.d
      Netra SPARC T4-1B              8.3.0.d
      SPARC T3-4                          8.3.0.d
      SPARC T3-2                          8.3.0.d
      SPARC T3-1B                        8.3.0.d
      SPARC T3-1                          8.3.0.d
      Sun Netra T5440 Server                        7.4.6.c
      Sun Netra T5220 Server                        7.4.6.c
      Sun SPARC Enterprise T5440 Server      7.4.6.c
      Sun SPARC Enterprise T5240 Server      7.4.6.c
      Sun SPARC Enterprise T5220 Server      7.4.6.c
      Sun SPARC Enterprise T5140 Server      7.4.6.c
      Sun SPARC Enterprise T5120 Server      7.4.6.c
      Sun Blade T6340 Server Module             7.4.6.c
      Sun Blade Netra T6340 Server Module    7.4.6.c
      Netra T6340 Server Module                    7.4.6.c
      Sun Blade T6320 Server Module             7.4.6.c

If both factors are present, then any reset of the Service Processor (SP) with the host running Solaris can result in corruption of memory and data loss.

Symptoms

The symptoms will vary depending on how the page that is corrupted is being used. Other data corruption is possible including user memory or files. The following two symptoms have been observed:

1) Applications may fail to start and/or core dump due to 'missing symbols' in linked libraries when the corruption occurs within the in-memory copy of the  library.

2) pkgchk(1M) may fail the checksum for a file.  If the file in question resides in-memory or within the filesystem cache, this potentially corrupt version will be used rather than the on-disk copy which remains intact and un-harmed.

Workaround

There is no workaround for this issue. While halting the OS before resetting the SP will prevent the problem, the SP will on occasion reset itself with little or no notice, thus the exposure cannot be eliminated entirely.

The following represents the minimum firmware patch and revision level for each platform that will resolve the issue:

Note that these are the minimum levels; any firmware patch and revision level greater than these will resolve the issue. ("...or later")

Patches

<SUNPATCH:17019082>,  <SUNPATCH:17199962>,  <SUNPATCH:17199954>,  <SUNPATCH:17199947>
<SUNPATCH:17199974>,  <SUNPATCH:150418-03>, <SUNPATCH:150419-03>, <SUNPATCH:150417-03>
<SUNPATCH:150415-03>, <SUNPATCH:150414-03>, <SUNPATCH:150416-03>, <SUNPATCH:150413-03>
<SUNPATCH:150411-03>, <SUNPATCH:150412-03>, <SUNPATCH:150409-03>, <SUNPATCH:150408-03>
<SUNPATCH:150410-03>, <SUNPATCH:150407-03>, <SUNPATCH:147313-09>, <SUNPATCH:147309-08>
<SUNPATCH:147311-07>, <SUNPATCH:147310-09>, <SUNPATCH:147307-09>, <SUNPATCH:147312-08>
<SUNPATCH:147314-08>, <SUNPATCH:147308-07>

History

27-Sep-2013: Document released; State: Resolved
01-Oct-2013: Additional clarification for Solaris patch info
04-Oct-2013: Added additional table of systems affected to "Occurrence" section
16-Oct-2013: Added minor formatting corrections for clarification

This regression was caused by the putback for 15491355 as delivered by patch
142909-17.  15491355 allows Solaris to reuse memory allocated to OBP.

There was a failure to verify that OBP was no longer utilizing all of the memory.
In this case an LDC channel was left open and a reset of the SP caused the SP to
attempt to open the channel again. The change closes the channel before handing
control over to Solaris.

Questions regarding this document should be addressed to
sunalertpublication_us_grp@oracle.com and copy the
responsible engineer/submitter below.

Internal Contributor/Submitter: David.Arneson@oracle.com
Internal Eng Responsible Engineer: David.Arneson@oracle.com
Oracle Knowledge Analyst: david.mariotto@oracle.com
Internal Eng Business Unit Group: Systems Group - SYS
Internal Escalation ID: 3-7680384601, 3-6895206011

References

This solution has no attachment