請協助解析一下ECC 記憶體的錯誤 - Linux

Table of Contents

最近Server在使用時出現類似下方的訊息
Message from syslogd@ at Sat Nov 29 20:19:28 2008 ...
luna kernel: EDAC MC0: UE row 7, channel-a= 2 channel-b= 3 labels "-":
(Branch=1 DRAM-Bank=2 RDWR=Write RAS=8 CAS=0 FATAL Err=0x4)
Message from syslogd@ at Sat Nov 29 20:19:31 2008 ...
luna kernel: EDAC MC0: UE row 3, channel-a= 2 channel-b= 3 labels "-":
(Branch=1 DRAM-Bank=1 RDWR=Read RAS=8180 CAS=0 FATAL Err=0x4)
Message from syslogd@ at Sat Nov 29 20:19:32 2008 ...
luna kernel: EDAC MC0: UE row 1, channel-a= 2 channel-b= 3 labels "-":
(Branch=1 DRAM-Bank=3 RDWR=Read RAS=5916 CAS=0 FATAL Err=0x4)
Message from syslogd@ at Sat Nov 29 20:19:33 2008 ...
luna kernel: EDAC MC0: UE row 7, channel-a= 2 channel-b= 3 labels "-":
(Branch=1 DRAM-Bank=2 RDWR=Write RAS=8 CAS=0 FATAL Err=0x4)

DRAM-Bank會顯示從0~3都有
主機板總共插了16條記憶體
排的順序如下

DIMM 4D
DIMM 4C
DIMM 4B
...
DIMM 1A

數字表示BANK(分1~4)
Branch分0,1 (BANK1,2 屬於Branch0, BANK3,4屬於Branch1)

我看了很久不知道錯誤訊息是否有對應到DIMM的位置
能請各位網友協助一下嗎?

基本上,錯誤訊息都是寫Branch=1,接著後面的DRAM-Bank會變
其他訊息大同小異

我找不到到底是錯哪條orz

--
thePainter.
◣◢
◤ ◣ http://www.wretch.cc/blog/myelf
◢ ◤ ◤ ◤ Wretch@BBS -> P_myelf
thePainter. φthePainter.

--

All Comments

Daniel avatarDaniel2008-12-07
比較笨的辦法直接用memtest 拔掉一半記憶體去測 沒問題就測
Sierra Rose avatarSierra Rose2008-12-12
另外一半 有問題就將原本的那一半再拔一半去測 總會找到 xD