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.

httpd causing <defunct> exim processes?

Discussion in 'General Discussion' started by mcentral, Jan 2, 2004.

  1. mcentral

    mcentral Registered

    Joined:
    Dec 1, 2003
    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Western North Carolina
    I have a 1.3 Celeron with RH 7.3 with a recently updated kernel, current stable version of CPanel and everything associated.

    What I would like to know is how is httpd tied in with Exim?

    The problem I am having is I have a server that serves a high load of email (several safe lists being run with a PHP script) daily. Even without the safe list script running, I still get a pile of <defunct> exim processes. During the sending and receiving of mail this server accumulates a high amount of <defunct> Exim processes. If I restart httpd the defunct exim processes (zombies) die. If I do nothing at all, this server will accumulate a couple of thousand <defunct> exim process within an hours time and will basically bring the server to a very slow crawl at some point. I know I can setup a cron to reset the httpd server every half hour (which I have done) BUT, for some unknown reason this causes the mail que to hold old messages that never get delivered. (I know, that seems very strange to me too) 2 days ago I found almost 20gigs of old email just sitting in the que. Yeah, almost 20 gigs. There has been several 1000 emails that have gone out successfully. I just can't find a pattern to this maddness!

    There are no mail-forms on this server. There is no spammer using my resources as I have monitored this server for weeks. I've tweaked, poked, probed, updated, removed and read every piece of info I could find about this issue and still have no resolve. I've done everything but give this thing an enima.

    This server has been like this since I first leased it. I attempted to repair it with a fresh update of CPanel. I even submitted a TT to RS to run a MEM Check on it. They say its ok. I'm at witts end and could use some help before I tie a rope around it and send it to HeadSurfer to be used as an anchor for his boat. If he has one.

    Again, If anyone knows how httpd is tied in with Exim (If in fact it is) I sure would like to know. Or any other ideas??

    Thanks
    Michael
     
  2. mcentral

    mcentral Registered

    Joined:
    Dec 1, 2003
    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Western North Carolina
    Any ideas anybody? CPanel Gurus? C'mon. Help a brother out!

    Or, am I stuck out here in the cold with a semi-broken box?
     
  3. mcentral

    mcentral Registered

    Joined:
    Dec 1, 2003
    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Western North Carolina
    ok. I'll post changes and hope I hit something.

    I'm not usually one that asks for help because I can usually fix things myself or one of you fantastic guru's out there posts a fix somewhere but this time I have a problem that seems very complicated yet is probably something extremely simple to fix I just don't know enough about programming of these particular programs to know where to look.

    Anybody feel free to jump in anytime.

    Changed the Timeout in httpd.conf - The number of seconds before receives and sends time out from default 300 to 100 seconds and immediately the <defunct> exim processes dropped from over 600 within 30 minutes to between 10 and 35 within 30 minutes. This does not appear to have had any effect on http processes.

    Load average also dropped from a constant 15 to 4 or less with a few spikes of course.
    Instead of the system holding more than 500 <defunct> (Zombie) processes it now only maintains less than 25 until a httpd reset via cron takes place every half hour and zeros that out.

    What I don't understand is why is Exim so effected by changes to the httpd.conf file?

    Will post more later
     
  4. mcentral

    mcentral Registered

    Joined:
    Dec 1, 2003
    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Western North Carolina
    I GIVE UP!

    Oh well. I give up and am going to trash can CPanel and go back to what I know works VERY well. This stinks.

    Disregard any previous posts as they didn't work out.

    Good luck.
     
  5. jek

    jek Member

    Joined:
    Jan 16, 2004
    Messages:
    8
    Likes Received:
    0
    Trophy Points:
    1
    ouch... I feel your pain...

    I'm getting tonnes of these... about 20 every minute.

    mysql 29801 0.0 0.0 0 0 pts/0 Z 17:44 0:00 [mysqld <defunct>]
    mysql 29876 0.0 0.0 0 0 pts/0 Z 17:45 0:00 [mysqld <defunct>]

    Wazzup cPanel gurus? Any ideas? I have to keep restarting the service to get rid of all this stuff.
     
  6. chirpy

    chirpy Well-Known Member

    Joined:
    Jun 15, 2002
    Messages:
    13,475
    Likes Received:
    20
    Trophy Points:
    38
    Location:
    Go on, have a guess
    Since one has nothing at all to do with the other, it won't, unless you have a general resource issue on your server, or someone is indeed using a flaw in a CGI/PHP script somewhere.

    Do you have plenty of free memory (cached or otherwise)?

    Are you getting very high iowait's? If so, there may be a bottle-neck on your disks.

    Have you gone over to the Exim mailing lists and asked them there? Since Exim is a third-party product, you're more likely to get help from the people who wrote it to such an issue.

    I'd ask the exact same questions about MySQL.
     
  7. jek

    jek Member

    Joined:
    Jan 16, 2004
    Messages:
    8
    Likes Received:
    0
    Trophy Points:
    1
    don't understand it...

    a mate who is a systems admin. solved my problem by restarting Apache.

    *scratches head*
     
  8. vabtz

    vabtz Member

    Joined:
    Jul 17, 2003
    Messages:
    7
    Likes Received:
    0
    Trophy Points:
    1
    Sorry man cpanel really is crumby on the support side. I have the same issue. I am going to migrate to software developed by a more professionally managed company too.

    Sucks too because my users love it.
     
  9. dave9000

    dave9000 Well-Known Member

    Joined:
    Apr 7, 2003
    Messages:
    891
    Likes Received:
    1
    Trophy Points:
    16
    Location:
    arkansas
    cPanel Access Level:
    Root Administrator
    Here is a scarey thought

    We are both developing the same thoughts about a lot of these so called system admins/ server owners

    Would hate to see what would happen if they had to manage all aspects of a NOC
     
  10. JBapt

    JBapt Active Member

    Joined:
    Feb 5, 2004
    Messages:
    41
    Likes Received:
    0
    Trophy Points:
    6
    Location:
    Porto - Portugal
    Hi Michael,

    first of all this is getting a bit tiresome. I cant see how you blame exim for a problem that can be caused by Apache, PHP, MySQL (if you are using it for the email databases), or even by your lame scripts.

    Also, i just cant see how you say that mails getting queued is strange for you. As you might not know an email message is a communication between two or more servers so its NORMAL and thank god it happens that messages sometimes get in queue for later delivery (just a wild thought but the destination server may be down, or any other kind of operation). Also there is a latency time sending mails so you must think you have a super computer that delivers mails instantly and goes one sending like a mad man. Another wild thought... read and understand the exim config file. There are a lot of bottlenecks and again thank god, or else, this would be spammers paradise.

    I know this wont help you at all, just to make you think and try to learn. If anybody hasnt answered you is because your problem is lame. CPanel support until now has given good support on real problems.
     
  11. kliest

    kliest Member

    Joined:
    Feb 23, 2004
    Messages:
    6
    Likes Received:
    0
    Trophy Points:
    1
    Hi,

    I have a simular problem. I also have a lot of scripts on my server that are only used to email. The server is mailing about 40,000 messages an hour. The problem I have is that the server loads is very high between 15 and 100. This is what I see when I type top:

    20:21:50 up 36 days, 13:05, 2 users, load average: 17.42, 23.22, 29.11
    292 processes: 233 sleeping, 8 running, 49 zombie, 2 stopped
    CPU0 states: 44.1% user 55.3% system 0.0% nice 0.0% iowait 0.0% idle
    CPU1 states: 9.3% user 89.1% system 0.0% nice 0.0% iowait 0.5% idle
    Mem: 1029760k av, 962884k used, 66876k free, 0k shrd, 141416k buff
    580068k actv, 106268k in_d, 18028k in_c
    Swap: 1052216k av, 180416k used, 871800k free 435532k cached

    PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND
    20104 mailnull 15 0 708 464 372 S 99.9 0.0 0:32 1 exim
    20997 mysql 15 0 37784 35M 1284 S 63.0 3.4 29:52 0 mysqld
    26922 root 15 0 6808 4120 3212 S 21.4 0.4 0:44 1 httpd
    24379 mailnull 20 0 0 0 0 Z 4.5 0.0 0:00 1 exim <defunct
    24386 root 20 0 3504 3504 2712 S 4.1 0.3 0:00 1 exim
    24370 mailnull 18 0 0 0 0 Z 3.1 0.0 0:00 1 exim <defunct
    24382 mailnull 19 0 3532 3532 2736 D 3.1 0.3 0:00 0 exim
    24397 root 21 0 2512 2512 2040 R 2.7 0.2 0:00 1 exim
    24366 mailnull 18 0 0 0 0 Z 2.5 0.0 0:00 0 exim <defunct
    24377 root 20 0 3460 3460 2684 S 2.5 0.3 0:00 1 exim
    24384 root 19 0 3496 3496 2716 S 2.5 0.3 0:00 1 exim
    24388 root 20 0 3492 3492 2712 S 2.5 0.3 0:00 1 exim
    24391 root 20 0 3500 3500 2712 S 2.1 0.3 0:00 1 exim
    24349 mailnull 16 0 0 0 0 Z 1.5 0.0 0:00 1 exim <defunct


    I have optimized the my.cnf as follows:

    [mysqld]
    set-variable = max_connections=500
    safe-show-database
    key_buffer = 16M
    myisam_sort_buffer_size = 64M
    join_buffer_size = 1M
    read_buffer_size = 1M
    table_cache = 1500
    thread_cache_size = 128
    interactive_timeout = 300
    wait_timeout = 10
    connect_timeout = 10
    max_allowed_packet = 16M
    max_connect_errors = 10
    query_cache_limit = 1M
    query_cache_size = 32M
    query_cache_type = 1
    skip-innodb

    [mysqld_safe]
    open_files_limit = 8192

    [mysqldump]
    quick
    max_allowed_packet = 16M

    [myisamchk]
    key_buffer = 256M
    sort_buffer = 256M
    read_buffer = 1M
    write_buffer = 1M

    Increasing the buffer size helped a little bit, but I still have high loads. At this moment I have setup a crontab to restart apache every 10 minutes. This is keeping the load down, but it is also causing that the sites are not loading when apache restarts, or when a lot of data needs to be loaded from a webpage. I have tried every fix I have found in this forum, but none of them helped.

    I have the following configuration:
    double pentium 2.8ghz processors
    1 gb RAM
    redhat 9
    whm 9.4.0

    Can anyone help me out with this.

    Edwin
     
Loading...

Share This Page