Results 1 to 14 of 14
  1. #1
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984

    Urgent Help Qmail Server

    My New Email server Dual Xeon P4 28 with adaptec 2410 card and sata 120 gb 3 hard disk seem running like a snail even for very small load, CPU is ok 90 plus % but load average 20+ cant figure out what is taking the load, I am running Clamav, SPamd etc even I disabled these the load also very high

  2. #2
    Join Date
    Mar 2003
    Location
    California USA
    Posts
    13,294
    is this a redhat enterprise box? is your iowait spiking?
    Steven Ciaburri | Industry's Best Server Management - Rack911.com
    Software Auditing - 400+ Vulnerabilities Found - Quote @ https://www.RACK911Labs.com
    Fully Managed Dedicated Servers (Las Vegas, New York City, & Amsterdam) (AS62710)
    FreeBSD & Linux Server Management, Security Auditing, Server Optimization, PCI Compliance

  3. #3
    Join Date
    Dec 2001
    Location
    Above The Clouds
    Posts
    6,999
    What's the output on:

    /var/qmail/bin/qmail-qstat
    Laurence Flynn @ atOmicVPS LTD
    Linux & Windows Cloud Hosting Solutions Powered by OnApp
    Fully Managed [Shared][Reseller][Cloud VPS] [Dedicated]
    Featuring the atOmicSTACK ● Speed ● Performance ● Reliability

  4. #4
    Join Date
    Apr 2001
    Location
    Palm Beach, FL
    Posts
    1,095
    What does your memory and swap look like? Also, what does 'top' show?
    Alex Llera
    Professional Server Management
    FreeBSD|Linux|HSphere|Cpanel|Plesk

  5. #5
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984
    17:14:47 up 1 day, 3:59, 3 users, load average: 25.06, 14.76, 13.68
    183 processes: 182 sleeping, 1 running, 0 zombie, 0 stopped
    CPU0 states: 1.0% user 2.4% system 0.0% nice 0.0% iowait 95.4% idle
    CPU1 states: 10.2% user 5.1% system 0.0% nice 0.0% iowait 84.0% idle
    CPU2 states: 6.2% user 1.4% system 0.0% nice 0.0% iowait 91.2% idle
    CPU3 states: 3.4% user 2.4% system 0.0% nice 0.0% iowait 93.0% idle
    Mem: 3098064k av, 3012840k used, 85224k free, 0k shrd, 470404k buff
    2191200k actv, 176764k in_d, 68840k in_c
    Swap: 2040212k av, 22520k used, 2017692k free 2073088k cached

  6. #6
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984
    /var/qmail/bin/qmail-qstat
    messages in queue: 662
    messages in queue but not yet preprocessed: 11

    Looks ok but is 10800 queuelifetime too long or too short? I have about 2 mbps throughput for this email server , really headache

  7. #7
    Join Date
    Dec 2001
    Location
    Above The Clouds
    Posts
    6,999
    Qmail looks fine. Check the MySQL monitor:

    watch -n 1 "uptime; free -m; mysqladmin -u admin --password=MYSQL_PASSWORD processlist "

    Plus output on ps -aux or ps -eafd - check for exploits.

    Concurrent apache processes: ps auwx | grep -c httpd
    Also: netstat -tan

    Look for misconfigured gateways via the neighbour table in arp: ip neigh|wc -l

    Possibilities are endless.
    Laurence Flynn @ atOmicVPS LTD
    Linux & Windows Cloud Hosting Solutions Powered by OnApp
    Fully Managed [Shared][Reseller][Cloud VPS] [Dedicated]
    Featuring the atOmicSTACK ● Speed ● Performance ● Reliability

  8. #8
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984
    Thanks for the tips. Even with load 10-20 is it ok? Sorry me used to low load......Also is queuelinetime of 10800 too little or too much? Sorri I am Bill Gates fan so trying to learn Linux seems a bit weird esp at my old age......please bear with me.......

  9. #9
    Join Date
    Dec 2001
    Location
    Above The Clouds
    Posts
    6,999
    Load of 10-20 is horrible. I don't even know what the queuelinetime variable is. Where did you find it? I'll see what one of our servers has...
    Laurence Flynn @ atOmicVPS LTD
    Linux & Windows Cloud Hosting Solutions Powered by OnApp
    Fully Managed [Shared][Reseller][Cloud VPS] [Dedicated]
    Featuring the atOmicSTACK ● Speed ● Performance ● Reliability

  10. #10
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984
    It is at /var/qmail/control queuelifetime.
    My headache traffic so low yet
    19:16:21 up 1 day, 6:01, 3 users, load average: 20.31, 17.56, 15.47
    149 processes: 148 sleeping, 1 running, 0 zombie, 0 stopped
    CPU0 states: 4.3% user 2.4% system 0.0% nice 0.0% iowait 92.1% idle
    CPU1 states: 3.1% user 1.1% system 0.0% nice 0.0% iowait 95.1% idle
    CPU2 states: 1.2% user 1.4% system 0.0% nice 0.0% iowait 96.2% idle
    CPU3 states: 20.0% user 3.3% system 0.0% nice 0.0% iowait 76.0% idle
    Mem: 3098064k av, 2999248k used, 98816k free, 0k shrd, 465900k buff
    2192488k actv, 192572k in_d, 65276k in_c
    Swap: 2040212k av, 22892k used, 2017320k free 2070472k cached

    The load is fantastic but CPU is ok, aspirin not enuff for me, looking for potassium cynaide soone

  11. #11
    Greetings:

    Queue life time is typically four to seven days. It is the time messages which are deferred (delayed) are kept until the mail server gives up trying to send them.

    I agree the server load is extremely high.

    Our mail server has the highest load of all our servers; and is typically between 0.75 and 1.50.

    1. Change your TCP session count to be at least 100 (if less than 100).

    2. Change your concurrencylocal to be at least 75 (if less than 75).

    3. Change your concurrencyremote to be at least 150 (if less than 150).

    NOTE on 2 and 3: You will need the qmail big concurrency patch if you ever need to set them higher than 240.

    Now, restart qmail.

    Also, if you see a lot of CNAME deferrals within your mail log file, it can be due to some providers DNS response and how qmail handles them.

    You can try to resolve this by applying the qmail big DNS patch.

    Thank you.
    ---
    Peter M. Abraham
    LinkedIn Profile

  12. #12
    Greetings:

    Two other, but related points:

    1. At http://www.qmail.org/top.html there are at least two qmail queue repair utilities (please note qmail must be stopped when you run them).

    You may want to use one to ensure your queue is in a correct state.

    2. There are several utilities which can help you clean up your queue in terms of MAILER-DAEMON and other related email which may be of no value.

    That might get your current queue down to a more manageable volume.

    Thank you.
    ---
    Peter M. Abraham
    LinkedIn Profile

  13. #13
    Join Date
    Mar 2004
    Location
    Singapore
    Posts
    6,984
    Thanks my expert brother God told me that it is due to the raid Adaptec Raid 2410 I am using Raid 5. He said some prob with the driver. Support for Redhat 9 is minimal, what should I do? Basically he is never wrong and he also cant do anything. Which means big trouble for me. Any other Demi Gods out there that can help? I am quite in deep ****.

  14. #14
    Join Date
    Jul 2001
    Location
    Singapore
    Posts
    1,790
    For queuelifetime and others, you can find your current value by issuing this command:
    /var/qmail/bin/qmail-showctl
    which will show you something like:
    queuelifetime: (Default.) Message lifetime in the queue is 604800 seconds.
    Standard is ONE WEEK which is 604800 seconds and if you want to lower or adjust that... simply create a filename /var/qmail/control/queuelifetime with your desired value then send a HUP to qmail.
    Whereby for your RAID... can you disable it or something... like can you remove the RAID to see whether do you have the same problem? This might be difficult since that server is already in production use I guess Or is your kernel from stock red hat or you compile your own? Just throw in couple of thoughts

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •