From: David Price (dprice@plugnpay.com)
Date: Wed Apr 16 2003 - 11:35:00 EDT
Partial Summary, Additional Questions
Thanks to all who responded.
Several responses pointed me to look into a "Famed" e-cache parity issue
that SUN experienced on there Enterprise line with CPU 400+ Mhz CPU's.
This "Best Practice Guide" describes it relatively well.
http://www.filibeto.org/sun/lib/hardware/enterprise_4500/BP_Ecache_10-16-01.
pdf
Symptoms of this error can cause the following error to appear in the
messages file:
Apr 4 23:50:41 vail unix: WARNING: [AFT1] EDP event on CPU7 Data access at
TL=0, errID 0x00000db5.ff543553
Apr 4 23:50:41 vail unix: AFSR 0x00000000.00408000 AFAR 0x00000000.4d7d3270
Apr 4 23:50:41 vail unix: AFSR.PSYND 0x8000(Score 95) AFSR.ETS 0x00 Fault_PC
0x6ed5a0
Apr 4 23:50:41 vail unix: UDBH 0x0000 UDBH.ESYND 0x00 UDBL 0x0000
UDBL.ESYND0x00
Apr 4 23:50:41 vail unix: [AFT2] errID
0x00000db5.ff543553PA=0x00000000.4d7d3270
Apr 4 23:50:41 vail unix: E$tag 0x00000000.0fc009af E$State: Modified
E$parity 0x07
The fix appears to be to replace the CPU's and/or CPU/RAM board. Newer
versions of the CPU are supposed to have the ecache mirrored to prevent
these types of crashes.
After closer examination of the Part numbers of the CPU's we had installed,
they appear to be of the newer version.
Part # 501-5816 which according to:
http://sunsolve.sun.com/handbook_pub/Devices/CPU_Module/UltraSPARC_464MHz_Ul
traII_Exx00.html
has the mirrored SRAM and mirrored Tag SRAM.
So I may be back to square 1.
We are still seeing an error message when running VTS which may or may not
be related.
FATAL mem: "read() at address 0x3fffffffff800000 [Board3, Bank0,
Size=2048MB, Intlv=2, MCTL=0x8541b09, MDEC=0x80001f8000000380: ] failed (Bad
address)."
Thanks again.
Dave
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:12 EDT