Results 1 to 1 of 1
hi guys, having a weird issue all of a sudden with my local linux server in my house. for the past couple weeks or so, the load will hit and ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
- 10-15-2007 #1
weird server load
having a weird issue all of a sudden with my local linux server in my house.
for the past couple weeks or so, the load will hit and stick around 2.0-5.0 and the cpu usage as presented by cacti constantly shows 20-30% cpu usage even when completely idle according to top.
i have not changed anything on this system other than adding 4 more machines to my local lan. one xp machine, and a wii/ps3/xbox 360. all 4 have been connected via dhcp...the rest of the lan has static ip's.
the server in question is running debian 4 and is running:
apache/mysql/snmpd for cacti
leafnode for local news group caching
bind9 for local dns caching
dhcpd3 for the 4 new machines
with shorewall the machine is acting as the router/firewall for lan->wan / wan->lan communications.
i did add ntop to monitor communications to estimate additional bandwidth, and this did seem to indeed cause a bit of lag in areas and additional load on the machine. the service has been shutdown and disabled from started up upon reboot. this fixed most of the issue, but the problem still exists.
the machine used to always have a load of 0.0 - 0.5 depending on what was being done other than file transfers off/to the machine via nfs and/or samba. this would naturally increase the load which still remained sane.
the machine specs are:
25gb hard drive
2x 10/100 nics (1 lite-on, 1 smc - rtl-8029)
i have attached below a image of the current cpu usage and load, along with the first few lines of top which show virtually 0 cpu usage which is not consistant with the load, nor the cpu usage graph as reported by cacti.
any help or ideas where to further troubleshoot this would be a great help. i've tried disabling pretty much all the services with no luck. im starting to just think all this is a bit to much for a 500mhz machine...but it shouldnt be.
top - 17:11:09 up 1 day, 11:00, 4 users, load average: 1.44, 1.55, 1.28 Tasks: 87 total, 2 running, 85 sleeping, 0 stopped, 0 zombie Cpu(s): 0.4% us, 2.7% sy, 0.0% ni, 0.0% id, 92.0% wa, 1.8% hi, 3.1% si Mem: 126908k total, 122984k used, 3924k free, 28408k buffers Swap: 979956k total, 29832k used, 950124k free, 24460k cached PID USER TIME+ %CPU %MEM PR NI S SWAP RES COMMAND 20379 chris 0:00.08 1.3 0.9 16 0 R 1044 1112 top 19856 root 0:00.90 0.9 0.0 15 0 S 0 0 pdflush 1 root 0:01.16 0.0 0.2 16 0 S 1228 264 init 2 root 0:00.00 0.0 0.0 34 19 S 0 0 ksoftirqd/0 3 root 0:00.02 0.0 0.0 5 -10 S 0 0 events/0 4 root 0:00.00 0.0 0.0 5 -10 S 0 0 khelper 15 root 0:14.51 0.0 0.0 5 -10 S 0 0 kblockd/0 38 root 0:00.00 0.0 0.0 5 -10 S 0 0 aio/0 37 root 1:37.39 0.0 0.0 15 0 S 0 0 kswapd0