Solaris9 Corrected Memory Error detected by CPU0 1

[marrow]

Aug 28 02:30:45 byctir02 AFSR 0x00000000.1000064a<FRC> AFAR 0x00000000.00000000 INVALID
Aug 28 02:30:45 byctir02 Fault_PC 0x1036088 Esynd 0x004a J_AID 3
Aug 28 02:30:45 byctir02 SUNW,UltraSPARC-IIIi: [ID 925616 kern.info] [AFT0] errID 0x0002c891.a5215938 Data Bit 15 was in error and corrected
Aug 28 02:30:45 byctir02 SUNW,UltraSPARC-IIIi: [ID 271662 kern.info] NOTICE: [AFT0] Corrected memory (FRC) Event detected by CPU0 at TL=0, errID 0x0002c891.a5215938
Aug 28 02:30:45 byctir02 AFSR 0x00000000.1000064a<FRC> AFAR 0x00000000.00000000 INVALID
Aug 28 02:30:45 byctir02 Fault_PC 0x1036088 Esynd 0x004a J_AID 3
Aug 28 02:30:45 byctir02 SUNW,UltraSPARC-IIIi: [ID 925616 kern.info] [AFT0] errID 0x0002c891.a5215938 Data Bit 15 was in error and corrected

This message is displayed in /var/adm/messages each night at 02:30 and sometimes at 14:30. Nothing in root crontab. Happened on a V440 and an old Ultra. Anybody seen this before? I'm told a patch ight fix

 

[Willemvrm]

Hi Marrow,

I guess that you see the errors at an interval of 12 hours is because the memory scrubbers runs every 12 hours. This mechanism reads the whole memory and this is why you see the errors, as there is something wrong on that memory location. But as you can see it is corrected, so will not cause any harm.
Depending on the OS kernel level, you can have memor page retirement which will mark the memory as bad, the os will not use that memory anymore, but even then the scrubber will report the error.
It could wel be that the page has already been retired.

On sunsolve.sun.com you can download the application cediag which will tel you exactly what is going on and if a dimm needs to be replaced.

Regards, Willem

 

[marrow]

Many Thanks for your reply Willemvrm, you have solved our mystery.