The Community Forums

Interact with an entire community of cPanel & WHM users!
  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

High server load

Discussion in 'General Discussion' started by GoWilkes, Jan 31, 2012.

  1. GoWilkes

    GoWilkes Well-Known Member

    Joined:
    Sep 26, 2006
    Messages:
    367
    Likes Received:
    1
    Trophy Points:
    18
    cPanel Access Level:
    Root Administrator
    I'm working with an older semi-managed server that hosts about 50 low-traffic websites.

    For the last several months, I'm seeing a very high server load of 2-15, even late at night. The management company is blowing it off since the high load isn't that noticeable, but IMO, it's not to be ignored; especially since I get regular reports that imap failed, exim failed, and occasionally that the IP is not responding to pings.

    I do run CSF firewall. Turning it off usually lowers the server load for a few minutes, but then it jumps back up after a few more minutes. So I don't think that it's necessarily causing the problem. I think?

    A scan for viruses with rkhunter comes back clean.

    Today, I received the following load report, showing a 15 minute average load of 3.89. Can you guys help me decipher what might be causing the high load?

    TIA,

    Jason


    Current Time: Tuesday, 31-Jan-2012 14:52:24 EST
    Restart Time: Wednesday, 28-Dec-2011 17:04:19 EST
    Parent Server Generation: 507
    Server uptime: 33 days 21 hours 48 minutes 5 seconds
    Total accesses: 2655942 - Total Traffic: 21.2 GB
    CPU Usage: u15.62 s3.4 cu14.64 cs0 - .00115% CPU load
    .907 requests/sec - 7.6 kB/second - 8.4 kB/request
    1 requests currently being processed, 9 idle workers


    *Edit note: I removed a bunch of pasted text that was impossible to read. - Jason*
     
    #1 GoWilkes, Jan 31, 2012
    Last edited: Feb 1, 2012
  2. JayFromEpic

    JayFromEpic Well-Known Member

    Joined:
    Apr 2, 2011
    Messages:
    204
    Likes Received:
    0
    Trophy Points:
    16
    Location:
    Scottsdale, AZ
    cPanel Access Level:
    Root Administrator
    Please don't post readouts like that, it makes it very difficult to read.

    If exim and IMAP are failing, there could be a lot of spam emails coming in or mass amounts of emails being sent out.

    Can you please check the cPanel accounts and see if the catchall addresses are on and if they are of decent size.
     
  3. GoWilkes

    GoWilkes Well-Known Member

    Joined:
    Sep 26, 2006
    Messages:
    367
    Likes Received:
    1
    Trophy Points:
    18
    cPanel Access Level:
    Root Administrator
    Sorry, it didn't look like that when I pasted it. It looks like the board here converted all of the extra whitespaces into ****.

    I looked at the Relayers list, and today there have been 704 emails sent, with the most from one account being 306. There are 67 in the mail queue. Does either of those seem like an exorbitant amount?

    I went through all of the default addresses, and almost all of them are set to :blackhole:. A few were set to fail and send a warning, so I changed them to :blackhole:, too. Of the ones that were forwarding to a default address on the server, the account looked fine.

    A few had 100-125MB in storage, though. That wouldn't be large enough to cause a problem, would it?

    FYI, after making the changes to the default address, the current server load is at 2.5. It's a little after midnight, so the traffic should be at the lowest, and most of the server updates are set to begin at 2am or later.

    Using top -c, the things currently using the highest CPU are:

    /usr/sbin/named -u named (around 8%)
    spamd child (constantly pops up for a second, then goes away; ranges from 10% to 45%)
    cpanellogd (popped up 3 times in 5 minutes; usage was around 50%)
     
    #3 GoWilkes, Jan 31, 2012
    Last edited: Jan 31, 2012
  4. alphawolf50

    alphawolf50 Well-Known Member

    Joined:
    Apr 28, 2011
    Messages:
    186
    Likes Received:
    2
    Trophy Points:
    18
    cPanel Access Level:
    Root Administrator
    Hi GoWilkes,

    Just a tip: if you put output inside CODE tags, it will retain its original formatting. Much easier to read that way :)
     
  5. GoWilkes

    GoWilkes Well-Known Member

    Joined:
    Sep 26, 2006
    Messages:
    367
    Likes Received:
    1
    Trophy Points:
    18
    cPanel Access Level:
    Root Administrator
    Thanks for the note!

    Still digging, I found this for yesterday under WHM's Daily Process Log:

    Code:
    User	Domain	% CPU	Process
    Top Processes
    root		70.2	/usr/bin/perl -MExtUtils::Command::MM -e pod2man -- --section3 --perm_rw644 (and a bunch more modules; 20 or 30 lines worth)
    named		7.5	/usr/sbin/named -u named
    named		7.4	/usr/sbin/named -u named
    root		63.0	cpanellogd - updating bandwidth for [acct]
    root		61.6	cpanellogd - updating bandwidth for [acct]
    mailnull		6.0	/usr/sbin/exim -bd -q60m
    mailnull		5.0	[exim]
    munin		46.0	/usr/bin/perl /usr/share/munin/munin-update
    mailnull		3.5	/usr/sbin/exim -bd -q60m
    mailman		24.0	/usr/local/bin/python2.4 -S /usr/local/cpanel/3rdparty/mailman/cron/checkdbs
    munin		24.0	/usr/bin/perl /usr/share/munin/munin-update
    munin		21.0	/usr/bin/perl /usr/share/munin/munin-update
    mailman		13.0	/usr/local/bin/python2.4 -S /usr/local/cpanel/3rdparty/mailman/cron/disabled
    
    That 70% for a bunch of Perl modules looks pretty ominous, and so does the 63 and 61.6% for cpanellogd. I also had a bunch of spamd child listed, but only one of those was large so I don't know if that's the hog or not.

    For today, the only high listing is named, which is still at 7.5% or so. I have a second server that's dedicated to another site, and it doesn't list named, cpanellogd, or the Perl modules as a top process.
     
Loading...

Share This Page