西西河

主题:求助:哪位老大知道Linux自动重起是咋回事? -- thunder

共:💬11 🌺1 新:
全看树展主题 · 分页首页 上页
/ 1
下页 末页
家园 求助:哪位老大知道Linux自动重起是咋回事?

Dual Opteron/SuSE 9.0 system, 今儿早上已经偷偷重起两回了。在DSL后面,router对这台机器只开了一个SSH port. 在第一次重起之后作了一次security update,大约过了一小时(?)听到硬盘磕磕响了两下,屏幕切换过去就见这家伙reboot。第一次重起也是听到硬盘在响。硬盘上东西应该都在,至少正工作的那个directory没问题。硬盘大概还有5.3G的空间。syslog说前几天有个家伙试图SSH登陆,mmd,什么人啊(219.???.???.???)闲得没事儿,还用root试了几次。这台机器用来计算的,没用它看过电影(呵呵,没装声卡,显卡on board),基本没用它上过网,平时上去就是SSH,cron每六个小时往自己yahoo email报一次IP。“always on”,没办法,现在shud down了。

家园 Linux我是外行。但一般而言,一个OS Reboot不外乎3个原因

1) Hardware problem. BIOS is not up-to-date, CPU is too hot; memory-timing setup is not right; disk cable is not connected very well and so forth.

2) Software problem. System file is corrupted, key initialization file is missing, version conflict, virus and so on.

3) Driver problem. Make sure you get the latest and greatest drivers.

Anyway, good luck!

家园 Thanks!!! But,

it was running ok for almost a year, and just in last week a program was running for 7 days occupying all the processors. it rebooted ok whenever a kernel update required it (last kernel required reboot was about a month ago). heard of virus can cause windows to reboot repeatedly. if this is what's happening to the Linux box ...

i'll check on the updates you suggested. thanks again!

家园 not familiar with suse

check /var/log/kern.log first.

You should provide more information.

can you get a login prompt?

what's your login prompt, X or text?

did you update kernel or libs such as libc recently?

Hope can help you.

家园 老大,i'm now quite shy for the following info

there's no /var/log/kern.log on suse, instead:

kernel 2.4.21-243-smp

/etc/sysconfig/kernel

INITRD_MODULES="scsi_mod sd_mod mptscsih reiserfs"

/var/log/warn

Oct 23 07:18:36 opteron kernel: bcm5700: eth1 NIC Link is DOWN

Oct 23 07:18:53 opteron kdm[1574]: fatal IO error 32 (Broken pipe)

/var/log/kdm.log

Failed to load module "v4l"

Could not init font path element /usr/X11R6/lib/X11/fonts/local, removing from list!

/usr/X11R6/lib/X11/fonts/kwintv

/usr/X11R6/lib/X11/fonts/CID

The eth0 is connected to the router, eth1 is blank. Is it the I/O error caused it? There was a dvd-rom in the drive, already removed it now. It boots ok. I can either work under text mode through ssh or in X on local. I did a kernel update like a month ago, it prompted to run lilo before reboot, followed their instructions. Booting is not a problem. netstat and "ps a" did not find suspicous staff.

Thank You!

家园 woops, there's two libc.a/libc.so

opteron:/var/log # ls -l /usr/lib64/libc.a

-rw-r--r-- 1 root root 19470534 2003-09-23 11:25 /usr/lib64/libc.a

opteron:/var/log # ls -l /usr/lib/libc.a

-rw-r--r-- 1 root root 13553180 2003-10-02 14:09 /usr/lib/libc.a

the .so's have the same time mark as the .a's

家园 老兄,俺也是个大菜鸟,

以下是我的一些菜鸟建议.

1. seems you have reiser patitions, try to unmount them. as I know, reiserfs is not stable.

2. turn off cron.

3. switch to text login prompt: change runlevel from 5 to 3 in /etc/inittab.

LIBC 应该没问题,一个是64位,一个是32位,不过文件的SIZE好象有点问题,太大了,我的也就3-4M的样子.

如果是被HACK了,那就帮不了你了.

家园 老大莫要谦虚,再教兄弟一把:

how to unmount reiser partition? i only have one swap and one linux native; i'll need a work-around if to turn off cron, to inform me the ip when i'm not at home.

will the run level have effect on the gui when i work locally?

it seems 64 bit machines do have larger libc. the opteron machine at office has a libc of 22M, and another itanium2 has a libc of 19M.

家园 不是谦虚,是真的很菜

又想到个方法.

光盘启动, DO SOME HARD WORKS ON THE COMPUTER, IF THE COMPUTER CRASHS AGAIN, IT'S HARDWARE PROBLEM.

BECAUSE ALL THE WORKING STUFFS ARE IN THE PARTITION, YOU CAN NOT UNMOUNT IT.

ACTUALLY MY IDEA IS TO ISOLATE THE BASIC SYSTEM FROM ALL OTHERS.

XF86 MAY HAVE SOME ODD PROBLEMS, SO TURN IT OFF. WHEN YOU CHANGE THE RUNLEVEL TO 3, OF COURSE YOU DON'T HAVE GUI AS YOU LOG ON. YOU CAN RUN STARTX TO GET IT BACK.

CRON ENABLES USERS TO EXECUTE COMMANDS AUTOMATICALLY AT A SPECIFIED TIME. AS YOU SAID, YOUR COMPUTER WILL REBOOT AFTER A WHILE, IT'S POSSIBLE THAT A PROGRAM RUN BY CRON CAUSE THIS CRASH, SO TURN IT OFF. TO TURN THE CRON OFF, IT DEPEND ON THE CRON DAEMON YOU ARE USING. IF YOU ARE USING VIXIE-CRON, RUN "/ETC/INIT.D/VIXIE-CRON STOP".

家园 THANK YOU SOOOO MUCH!

will google out boot-from-cd, it's been taking you a lot time. thanks! already copied down those with the run level, should be ok as it's been up for a while. :D let's pretend the cron is safe for the moment :p

again,

家园 YOU ARE WELCOME
全看树展主题 · 分页首页 上页
/ 1
下页 末页


有趣有益,互惠互利;开阔视野,博采众长。
虚拟的网络,真实的人。天南地北客,相逢皆朋友

Copyright © cchere 西西河