I recently upgraded my box to Redhat 7.2. It's K6-2 400Mhz / 256MB RAM / 2x6 GB IDE HDD. After the upgrade problems appeared. Every few hours the server locks :(. It does respond to pings and open connections to ports but no answers such as login prompt are displayed. Before the upgrade I didn't have such problems. There is nothing in the logs. I started "procinfo -d -n 30" and noticed there is something wrong with the load average.

Short before:

Linux 2.4.18i386 (root@72builder) (gcc 2.96 20000731 ) #1 1CPU [test]

Memory: Total Used Free Shared Buffers Cached
Mem: 0 206 -206 0 4 7
Swap: 0 0 0

Bootup: Sat Jul 13 16:40:56 2002 Load average: 0.38 0.49 0.51 1/129 5634

user :0:00:07.61 25.4% page in :180 disk 1:22r67w
nice : 0:00:00.00 0.0% page out:1750
system:0:00:01.49 5.0% swap in :0
idle : 0:00:20.90 69.7% swap out: 0
uptime:17:21:35.19 context : 3181

irq 0:3000 timer irq 8: 0 rtc
irq 1: 0 keyboard irq 9: 4808 eth0
irq 2: 0 cascade [4]irq 14:1004 ide0
irq 4:0 serial irq 15:2856 ide1


and 'top':


10:06am up 17:25, 4 users, load average: 0.84, 0.58, 0.53
141 processes: 137 sleeping, 4 running, 0 zombie, 0 stopped
CPU states: 27.8% user, 5.8% system, 0.0% nice, 66.2% idle
Mem: 256116K av, 249688K used, 6428K free,0K shrd, 15916K buff
Swap: 524464K av, 28088K used, 496376K free 76420K cached.

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
5869 nobody 9 0 5396 4904 2344 S 5.5 1.9 0:01 httpd
2686 nobody 9 0 5524 5260 2752 S 4.4 2.0 0:10 httpd
2578 nobody 9 0 5600 5452 2868 S 4.2 2.1 0:20 httpd
2576 nobody 9 0 6048 5848 2820 S 2.2 2.2 0:11 httpd
2952 nobody 15 0 6304 6136 2860 R 2.2 2.3 0:20 httpd
3282 nobody 9 0 5628 5452 2816 S 2.0 2.1 0:10 httpd
1704 nobody 9 0 5272 5104 2836 S 1.7 1.9 0:21 httpd
3284 nobody 10 0 5452 5276 2840 S 1.7 2.0 0:14 httpd
5871 mysql 9 0 43196 19M 1864 S 1.4 7.8 0:00 mysqld
2684 mysql 9 0 43196 19M 1864 S 0.5 7.8 0:01 mysqld
3658 nobody 10 0 5704 5440 2756 S 0.5 2.1 0:11 httpd


And two minutes later (the server didn't respond to web request already):


Linux 2.4.18i386 (root@72builder) (gcc 2.96 20000731 ) #1 1CPU [test]

Memory:Total Used Free Shared Buffers Cached
Mem: 0 0 0 0 0 0
Swap: 0 0 0

Bootup: Sat Jul 13 16:40:56 2002 Load average: 119.84 110.66 78.51 1/234 5986


user : 0:00:00.00 0.0% page in : 0 disk 1: 0r 0w
nice : 0:00:00.00 0.0% page out: 0
system: 0:00:00.04 0.1% swap in : 0
idle : 0:00:29.96 99.9% swap out: 0
uptime: 17:48:05.19 context : 1360

irq 0: 3000 timer irq 8: 0 rtc
irq 1: 153 keyboard irq 9: 5115 eth0
irq 2: 0 cascade [4] irq 14: 0 ide0
irq 4: 0 serial irq 15: 0 ide1


I was unable to execute 'top'

The server load is extremly high and the cpu usage is only 0.01%. The only solution I found was to reboot the server.

What could be the reason of such behavior? :confused:

Eish, that's a tough call unless you know what's running and using all the resources.

Do you know how to use screen? In ssh, open a screen with a window running top so you can go back to it when the loads go up and see what's causing the problems.

looks like maybe someone had an out of control php page or something running.
probably got caught in a loop and just stayed there.

just a thought

Originally posted by Joey
Do you know how to use screen? In ssh, open a screen with a window running top so you can go back to it when the loads go up and see what's causing the problems.

That is exactly what I did. But only procinfo was running and top which I had in other screen session locked (short before the load average went up) :(

I'm trying again to log 'top' and 'procinfo'. Maybe this time the top will not lock.

It happened again but again I was not able to check what processes are running. I saw only that the load goes high very fast. I don't think it could be caused by any user application because the procinfo showed that the cpu usage is less than 0.2%. Could it be a hardware issue?