LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - General (http://www.linuxquestions.org/questions/linux-general-1/)
-   -   MY server keeps crashing and I don't know why... (http://www.linuxquestions.org/questions/linux-general-1/my-server-keeps-crashing-and-i-dont-know-why-70212/)

Electrode 07-05-2003 12:35 AM

MY server keeps crashing and I don't know why...
 
This has been happening for 2 days now. I will be browsing a forum, talking on IRC, or just any of my normal activities, when suddenly my internet connection goes dead. I switch my KVM over to my server and find that it is completely frozen solid. No error messages, just totally frozen.

What makes this as frustrating as it is is that I haven't the slightest idea what the cause of the freezing is. I can't think of anything I've done to the server in the days leading up to the start of the failures that would cause it. I know that the CPU isn't overheating, since I can compile code for hours without a hiccup.

Specs:
Pentium 3 866
Asus CUSL2 mobo
512mb RAM
hda: WD 80GB; hdb and hdc: Maxtor 120GB; hdd and sr0: HP CD-RW
Slackware 9.0
Kernel 2.4.21-ac4 (was running -rc8-ac1 before today)
eth0: 3Com 10mbps; eth1: Realtek 100mbps
Running IPMASQ

Any ideas? I'm at a dead end and I'm already sick of resetting my server every few hours.

Electro 07-05-2003 04:33 AM

It could be your ISP has change their setup procedures. Try using automatically get IP address on the NIC that connects to the modem or internet connection. Check your logs. Also change your memory type to ECC or registered DIMMS. Check temperature on hard drives. You can add a fan to the hard drives if the hard drives are near maximum temperature readings. Use both an In-line UPS and a line conditioner if possible. Check voltage readings from ATX power supply. if the voltage levels are low, get another power supply that has a higher output power.

Electrode 07-05-2003 02:19 PM

I mounted the hard drives in coolers after I had a drive overheat and take a bunch of data along with it. :)

I don't really have the money to buy ECC RAM or a mobo that supports it. Nor do I have the money for a UPS. The external NIC has always used DHCP. The PSU is producing more than enough power.

There is only one thing that seems consistant about these freezes: I'm always transmitting to a remote host from a box behind NAT when it happens. Sometimes it will happen when saying something in an IRC channel, sometimes it happens when sending a request for a web page, it happened once while playing Quake 3. It has never happened while I was away.

david_ross 07-05-2003 02:22 PM

Take a look at the message log to see if there were any events at the time of the crash:
/var/log/messages

If you are worried about overheating and have a motherboard with sensors then take a look at gkrellm:
http://www.gkrellm.net

Electrode 07-05-2003 02:34 PM

no messages in /var/log/messages.

It hasn't crashed since yesterday, so perhaps the problem has resolved itself. For all I know, I might have been under attack or something...

Electrode 07-06-2003 12:31 AM

Hmm, it happened again just now. While getting my email messages, it locked up, and when I rebooted I found my root disk to be very "fscked up". Barely enough survived to boot. This is probably because the filesystem happened to be committing when the crash occured.

So, anyone else have ideas? I don't want this to happen again.

david_ross 07-06-2003 10:53 AM

Try running "top" and looking for proccesses that are using a lot of the cpu or a lot of memory. It could be that a program is constantly eating up memory. You can use M & P to sort by proccessor and memory usage. You may want to leave this runnign and just switch back to it every now and again to see if things have changed drastically.


All times are GMT -5. The time now is 10:53 PM.