High system load

[zweb]

My mail server seem to slow down drastically recently.. it sure aint the bandwidth problem and the top process as follow..

14:48:48 up 2:19, 1 user, load average: 5.88, 4.92, 5.20
97 processes: 90 sleeping, 6 running, 1 zombie, 0 stopped
CPU states: 15.7% user 83.2% system 0.7% nice 0.0% iowait 0.1% idle
Mem: 513852k av, 487752k used, 26100k free, 0k shrd, 42336k buff
350188k actv, 11804k in_d, 10268k in_c
Swap: 1052248k av, 17352k used, 1034896k free 237908k cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
1991 apache 16 0 12604 11M 2508 S 7.3 2.3 0:00 0 httpd
29517 root 25 0 2720 1992 1460 D 6.5 0.3 0:09 0 exim
28568 root 25 0 2716 2716 1460 S 4.9 0.5 0:10 0 exim
24028 root 25 0 2720 2720 1460 S 3.1 0.5 1:10 0 exim
17079 root 25 0 2720 2720 1460 D 0.7 0.5 1:11 0 exim
4880 root 25 0 2748 2748 1484 R 0.5 0.5 0:00 0 exim
1926 apache 15 0 13096 12M 2608 S 0.1 2.4 0:03 0 httpd
2004 root 15 0 1096 1096 868 R 0.1 0.2 0:00 0 top
4606 wizwerx1 24 0 0 0 0 Z 0.1 0.0 0:00 0 exim <defunct>
1 root 15 0 108 84 56 S 0.0 0.0 0:04 0 init
2 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 keventd
3 root 34 19 0 0 0 SWN 0.0 0.0 0:00 0 ksoftirqd_CPU0
8 root 25 0 0 0 0 SW 0.0 0.0 0:00 0 bdflush
4 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kswapd
5 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kscand/DMA
6 root 15 0 0 0 0 SW 0.0 0.0 0:07 0 kscand/Normal
7 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kscand/HighMem
9 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kupdated
10 root 25 0 0 0 0 SW 0.0 0.0 0:00 0 mdrecoveryd
18 root 15 0 0 0 0 SW 0.0 0.0 0:03 0 kjournald
603 root 15 0 0 0 0 SW 0.0 0.0 0:00 0 kjournald
1071 root 15 0 240 212 156 R 0.0 0.0 0:01 0 syslogd
1075 root 15 0 176 164 120 S 0.0 0.0 0:00 0 klogd
1112 rpcuser 25 0 80 0 0 SW 0.0 0.0 0:00 0 rpc.statd
1688 root 16 0 480 348 236 S 0.0 0.0 0:01 0 sshd
1702 root 15 0 412 368 288 S 0.0 0.0 0:00 0 xinetd
1711 clamav 15 0 7176 7156 696 S 0.0 1.3 0:51 0 clamd
1720 root 15 0 244 232 208 S 0.0 0.0 0:02 0 da-popb4smtp
1729 nobody 16 0 312 76 64 S 0.0 0.0 0:00 0 directadmin
1777 root 25 0 56 4 0 S 0.0 0.0 0:00 0 gpm
1853 root 15 0 6792 6064 60 S 0.0 1.1 0:00 0 httpd
1872 ftp 15 0 500 500 116 S 0.0 0.0 0:00 0 proftpd
1889 root 15 0 96 96 24 S 0.0 0.0 0:00 0 vm-pop3d
1910 root 15 0 412 412 336 S 0.0 0.0 0:00 0 crond
1922 apache 15 0 12628 11M 1996 S 0.0 2.3 0:06 0 httpd
1923 apache 15 0 13576 12M 2736 S 0.0 2.5 0:07 0 httpd
1924 apache 15 0 12520 11M 2172 S 0.0 2.3 0:05 0 httpd
1925 apache 15 0 12412 11M 2068 S 0.0 2.3 0:06 0 httpd
1927 apache 15 0 12140 11M 1956 S 0.0 2.2 0:03 0 httpd
1928 apache 15 0 12096 11M 1972 S 0.0 2.2 0:04 0 httpd
1929 apache 15 0 12748 11M 1996 S 0.0 2.3 0:09 0 httpd



Please advice on the high system load.. running on RH 9 p3 800 cpu with 512sdram

 

[zweb]

Ok.. after searching through exim list.. found out that i actually have too much exim process running.. some are stuck and strangly is run as root and not mail... i've entered the following into my exim.conf for thoese who have the same probelm as me..

# How much load we're willing to accept
# At this LA only hosts in smtp_reserve_hosts can connect
smtp_load_reserve = 5
# When to abandon queue runs
deliver_queue_load_max = 5
# After a this load we only queue incoming messages
queue_only_load = 5
# max queue running processes to run simultaneously
queue_run_max = 7
# Lets you set how many SMTP calls are handled
# This allows many more TCP connections in wait state
smtp_connect_backlog = 50
# When we start just queueing
smtp_accept_queue = 15
# We will not accpet more than 200 (changed to 20)
# incoming SMTP calls at once
smtp_accept_max = 30


now my load is back to < 1 avrg.
dont know if this is the permenant solution..

 

[nobaloney]

It should solve the problem permanently, although numbers may have to be tweaked from time to time.

At a cost of a slowdown in accepting email.

However you may not notice that.

Do you get a lot of spam on your server?

In the long run the best way to solve the spam problem is to use SpamBlocker, which is built into DA.

Jeff

 

[zweb]

yup i do have alot of spams... some dictonary attacks which are block by spam blocker.. perhaps its just the huge amount of mails which my clients are getting.. but anyhow.. i'm putting in another cpu and 512 ram to see how it goes.. seems like some crondjob php script which using sendmail function is causing to load to be high.. i'm still monitoring it..

 

[zweb]

ok just assuming here.. everytime there is a syncronise error in exim log or some unexpected error or lost connection, an exim program running as root will pop up which will hog most of my cpu ranging from 2secs to 30min... is this normal?

 

[nobaloney]

What does the line look like in your ps or top output?

Questions such as this would be best answered in the exim-users mailing list but if you post it here you may get a response.

Jeff