隨機伺服器重啟

Jan  6 22:43:27 HOSTkernel: [34400.406538] Firewall: *Port Flood* IN=eth0 OUT= MAC=00:25:90:33:2c:d4:00:12:7f:34:7d:bf:08:00 SRC=141.0.9.77 DST=SERVER IP LEN=60 TOS=0x00 PREC=0x00 TTL=59 ID=2055 DF PROTO=TCP SPT=39403 DPT=80 WINDOW=5840 RES=0x00 SYN URGP=0
Jan  6 22:43:27 HOSTkernel: [34400.407073] Firewall: *Port Flood* IN=eth0 OUT= MAC=00:25:90:33:2c:d4:00:12:7f:34:7d:bf:08:00 SRC=141.0.9.77 DST=SERVER IP LEN=60 TOS=0x00 PREC=0x00 TTL=59 ID=37445 DF PROTO=TCP SPT=39404 DPT=80 WINDOW=5840 RES=0x00 SYN URGP=0
Jan  6 22:44:43 HOSTpure-ftpd: (?@127.0.0.1) [INFO] New connection from 127.0.0.1
Jan  6 22:44:44 HOSTpure-ftpd: (?@127.0.0.1) [INFO] __cpanel__service__auth__ftpd__CAWArGbwUe_gSaVFaB3LcPVXMtxsvsPLH5aO5Ps5aTHj1G7UbWKZwETT5mjV46YP is now logged in
Jan  6 22:44:44 HOSTpure-ftpd: (__cpanel__service__auth__ftpd__CAWArGbwUe_gSaVFaB3LcPVXMtxsvsPLH5aO5Ps5aTHj1G7UbWKZwETT5mjV46YP@127.0.0.1) [INFO] Logout.
Jan  6 22:45:52 HOSTkernel: [34545.899779] Firewall: *ICMP_IN Blocked* IN=eth0 OUT= MAC=00:25:90:33:2c:d4:00:12:7f:34:7d:bf:08:00 SRC=183.23.22.109 DST=SERVER IP LEN=28 TOS=0x00 PREC=0x00 TTL=113 ID=9188 PROTO=ICMP TYPE=8 CODE=0 ID=299 SEQ=45619
Jan  6 18:00:55 HOSTkernel: imklog 5.8.10, log source = /proc/kmsg started.
Jan  6 18:00:55 HOSTrsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1779" x-info="http://www.rsyslog.com"] start
Jan  6 18:00:55 HOSTkernel: [    0.000000] Initializing cgroup subsys cpuset
Jan  6 18:00:55 HOSTkernel: [    0.000000] Initializing cgroup subsys cpu
Jan  6 18:00:55 HOSTkernel: [    0.000000] Linux version 2.6.32-379.14.1.lve1.1.9.9.el6.x86_64 (mockbuild@koji.cloudlinux.com) (gcc version 4.4.6 20120305 (Red Hat 4.4.6-4) (GCC) ) #1 SMP Thu Dec 6 07:12:24 EST 2012
Jan  6 18:00:55 HOSTkernel: [    0.000000] Command line: ro root=UUID=f5be9b96-da41-49fd-a8e2-dbbf37f5cd7d rd_NO_LUKS rd_NO_LVM LANG=en_US.UTF-8 rd_NO_MD quiet SYSFONT=latarcyrheb-sun16 rhgb crashkernel=auto  KEYBOARDTYPE=pc KEYTABLE=us rd_NO_DM panic=5 processor.max_cstate=0 nohz=off
Jan  6 18:00:55 HOSTkernel: [    0.000000] KERNEL supported cpus:
Jan  6 18:00:55 HOSTkernel: [    0.000000]   Intel GenuineIntel
Jan  6 18:00:55 HOSTkernel: [    0.000000]   AMD AuthenticAMD
Jan  6 18:00:55 HOSTkernel: [    0.000000]   Centaur CentaurHauls
Jan  6 18:00:55 HOSTkernel: [    0.000000] BIOS-provided physical RAM map:
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 0000000000000000 - 000000000009d800 (usable)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 000000000009d800 - 00000000000a0000 (reserved)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 00000000000e0000 - 0000000000100000 (reserved)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 0000000000100000 - 00000000bf7a0000 (usable)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 00000000bf7a0000 - 00000000bf7ae000 (ACPI data)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 00000000bf7ae000 - 00000000bf7d0000 (ACPI NVS)
Jan  6 18:00:55 HOSTkernel: [    0.000000]  BIOS-e820: 00000000bf7d0000 - 00000000bf7e0000 (reserved)

然後最後它會在正確的時間恢復。奇怪的？

…

確實，今天伺服器已經隨機重啟了兩次，出乎意料。幾週來我沒有遇到任何問題。最近，將驅動器 DDRescue 複製到新驅動器，執行 fsck 一切正常，這就是一切開始的時候。

想法？

uname -a Linux SERVER 2.6.32-379.14.1.lve1.1.9.9.el6.x86_64 #1 SMP Thu Dec 6 07:12:24 EST 2012 x86_64 x86_64 x86_64 GNU/Linux

您已經 fscked，大概不是驅動器。這是記憶體出現故障時所期望的確切行為。您可以使用 memtest86 或memtest86+ （在UBCD上傳入？）並執行幾個記憶體週期。顯然，對於生產伺服器，最好交換記憶體並在測試機器上執行它。

您的伺服器可能有一個壞的 RAM。使用 memtest 對其進行測試。也看看這個文章。

引用自：https://serverfault.com/questions/463587