CentOS crash, like two times a day =[

[fishown]

Hi ther im new =],
Im faceing a frastrated problem,
My box crash tow times a day, its started from once in a week until now that its once or twise a day!.
Im kinda newB and i dont know where to look to even find whats cause this,
so i came here to ask you guys for help,

Tech spec:

OS: centOS 5.6( updated thought yum from 5.5)
Proccessor: Intel xeon x3450b quadcore 2.66 ( shows 8 cores on logs, wierd)
Ram: 8Gm (dont remember the brand, if its critical i will update)
HDD: WD 2TB Black (x2)

anything else? , hoo and i used "The perfect server centos 5.5" tutorial on howtoforge.com.

thank you =].

[Irithori]

Define "crash".

Generally I would look into dmesg, /var/log/messages, application logs, statusreport of e.g. raidcontroller userland tools.
Also check if the disk is full, check your monitoring system for e.g. swapping.
Does it happen at the same time? Then maybe a cronjob does something nasty.

[fishown]

Define "crash".

Generally I would look into dmesg, /var/log/messages, application logs, statusreport of e.g. raidcontroller userland tools.
Also check if the disk is full, check your monitoring system for e.g. swapping.
Does it happen at the same time? Then maybe a cronjob does something nasty.

you right i did forgot to define "crash"
so, by crash i ment: black screen, just freeze.
cron: nope i have just three crons that run once a min. and no it dosend crash on the
same time or same period.
disc-usage: nope 73% free (out of 4tb)
var/log/messages: alredy looked nothing in ther, just stop the log until i reboot

and alredy looked into dmegs, and i dont know what to look for T_T.
and thouse tools you metions the same, im kinda new for linux, just 2.5 month since my first expirience.
if you can get into detail

[fishown]

Ive loged the system status just befor the crash....
if some1 see something wrong please tell me.


____________________________________data block_________________________________________
16:17:01 up 1 day, 19:04, 2 users, load average: 3.69, 3.51, 3.15

Linux 2.6.18-238.9.1.el5 (cssc.co.il) 05/19/11 16:17:01
CPU %user %nice %system %iowait %steal %idle 16:17:02
all 27.68 0.00 5.74 0.25 0.00 66.33
Average: all 27.68 0.00 5.74 0.25 0.00 66.33

total used free shared buffers cached Mem:
7974 7749 224 0 137 5227
-/+ buffers/cache:
2384 5590
Swap: 9983 0 9983

______________________________end of data block ___________________________________________

____________________________________data block_________________________________________
16:18:01 up 1 day, 19:05, 2 users, load average: 3.35, 3.43, 3.14

Linux 2.6.18-238.9.1.el5 (cssc.co.il) 05/19/11 16:18:01
CPU %user %nice %system %iowait %steal %idle 16:18:02
all 32.17 0.00 6.63 0.00 0.00 61.20
Average: all 32.17 0.00 6.63 0.00 0.00 61.20

total used free shared buffers cached Mem:
7974 7764 210 0 138 5238
-/+ buffers/cache:
2387 5587
Swap: 9983 0 9983

______________________________end of data block ___________________________________________

[atreyu]

are you running in graphical/desktop mode, or console only? i've had lock-ups a lot in CentOS/Fedora w/buggy video drivers. sometimes adding nomodeset at boot time help, too.

[fishown]

are you running in graphical/desktop mode, or console only? i've had lock-ups a lot in CentOS/Fedora w/buggy video drivers. sometimes adding nomodeset at boot time help, too.

nope, im running console only...
thanks for the itrest, any ideas?

By the way, I rad this post says that its might be the ram,
I used memtest86+ ....0 errors,
but this guy had 0 erros as well, he replaiced his ram, and walla...
do you think its related?

[atreyu]

yes, memory can definitely cause grief. before replacing memory, though, just pull out one stick at a time (if your machine allows that) and try them in different slots, too.

[fishown]

yes, memory can definitely cause grief. before replacing memory, though, just pull out one stick at a time (if your machine allows that) and try them in different slots, too.

hi ther,
so today im gonna update the kernel, see if something will change,
In case of ofrther crashes, i will pop out the ram and inform

[MikeTbob]

Try disk check also.
You might need some help with this command but basically you want to boot from a LiveCD and run the fsck test on your hard drives. They may not be failing but they could have errors that would cause this.

[Rubberman]

I've had overheating memory cause this sort of problem. RHEL/CentOS 5 doesn't have as good support for system thermal sensors that RHEL/SL 6 does. Once I upgraded to Scientific Linux (SL) 6, I found the problem, adjusted the slots I had the RAM plugged into for better air flow, and the problem went away, especially since the temperature of the RAM dropped 25-30 degrees.