求助/var/log/mcelog文件达到40G的原因

20032007

dell7910工作站，redhat 6.5操作系统，

[root@jj log]# df
Filesystem                   1K-blocks      Used Available Use% Mounted on
/dev/mapper/vg_jj-lv_root   51606140  51589756         0 100% /
tmpfs                         16499048        72  16498976   1% /dev/shm
/dev/sda1                       495844     41321    428923   9% /boot
/dev/mapper/vg_jj-lv_home 1854530800 777109956 983216124  45% /home

仔细检查后发现/var/log下的有个mcelog文件，大小达到39.8G，用cat /dev/null > /var/log/mcelog
后，能恢复正常，但是开机几天后，这个文件有变成了39.8G，导致上面的应用总是报错

切换到root用户下，mcelog显示：

Hardware event. This is not a software error.
MCE 29
CPU 8 BANK 7
MISC 42184000 ADDR 49493c540
TIME 1511174096 Mon Nov 20 18:34:56 2017
MCG status:
MCi status:
Error overflow
Corrected error
MCi_MISC register valid
MCi_ADDR register valid
MCA: MEMORY CONTROLLER RD_CHANNEL0_ERR
Transaction: Memory read error
STATUS cc00008000010090 MCGSTATUS 0
MCGCAP 7000c16 APICID 10 SOCKETID 1
CPUID Vendor Intel Family 6 Model 63
Hardware event. This is not a software error.
MCE 30
CPU 8 BANK 7
MISC 526a2a00 ADDR 49493c800
TIME 1511174096 Mon Nov 20 18:34:56 2017
MCG status:
MCi status:
Error overflow
Corrected error
MCi_MISC register valid
MCi_ADDR register valid
MCA: MEMORY CONTROLLER RD_CHANNEL0_ERR
Transaction: Memory read error
STATUS cc00008000010090 MCGSTATUS 0
MCGCAP 7000c16 APICID 10 SOCKETID 1
CPUID Vendor Intel Family 6 Model 63
Hardware event. This is not a software error.
MCE 31
CPU 8 BANK 7
MISC 4238a200 ADDR 46e5dda80
TIME 1511174096 Mon Nov 20 18:34:56 2017
MCG status:
MCi status:
Corrected error
MCi_MISC register valid
MCi_ADDR register valid
MCA: MEMORY CONTROLLER RD_CHANNEL0_ERR
Transaction: Memory read error
STATUS 8c00004000010090 MCGSTATUS 0
MCGCAP 7000c16 APICID 10 SOCKETID 1
CPUID Vendor Intel Family 6 Model 63


求教这是什么问题，如何解决，感谢各位专家指点。。。

20032007

基本判断为内存问题，待解决中。。。。。。。。。

q1208c

回复 1# 20032007

有一条内存坏了。ECC纠错成功。
建议更换内存。或者可以暂关掉mcelog。