ECC Memory: EDAC module error

[RAID-5]

Hi, I install CentOS 5 on an i82875p chipset based motherboard yesterday and I always get EDAC Unrecoverable Error message on the console.

Sep 12 09:31:46 charlie kernel: EDAC MC0: UE page 0x37773, offset 0x0, grain 4096, row 1, labels ":": i82875p UE
Sep 12 09:31:47 charlie kernel: EDAC MC0: UE page 0x374f3, offset 0x0, grain 4096, row 1, labels ":": i82875p UE
Sep 12 09:31:48 charlie kernel: EDAC MC0: UE page 0x3767b, offset 0x0, grain 4096, row 1, labels ":": i82875p UE

This is not an hardware error... it seems to be false errors. I try to disable ECC on the BIOS (no error is shown) and then add this to modprobe.conf;

alias i82875p_edac /dev/null
alias edac_mc /dev/null
options edac_mc panic_on_ue=0

After reboot, I enable ECC in BIOS, but error message occur again. Any idea to solve this?

Thanks.

[foxb]

https://bugzilla.redhat.com/show_bug.cgi?id=191506

[RAID-5]

Sorry for slow reply. I already read this bug report. For my case, it seems to be an issue with i82875p chipset... I got a lot of machines running on Intel 7xxx chipset without any problem.

Any suggestions to solve this?

thank you.

[ebuck]

Assuming everything works, you can change the level at which warning messages are sent to the console so that at least you can get work done (and ignore these messages). Use sysctl to change the kernel printk setting.

Ed

[RAID-5]

I solve my case by adding "alias i82875p_edac panic_on_ue=0" to the following lines in modprobe.conf


alias i82875p_edac /dev/null
alias edac_mc /dev/null
options edac_mc panic_on_ue=0


Thanks for you're suggestions.