CPU usage too high, SMTP status errors.

Issue #127 resolved
Emilio Martin Lundgaard Lopez created an issue

Hi, I've set up Piler but I run into the issue that my CPU usage is too high (even if I haven't indexed any mails yet) and I get some STMP status errors. For example right now I get

CPU usage: 92%
SMTP status:
    piler: ERROR
    smarthost: ERROR

I noticed this similar issue but none of the information there really helped me. I'm runnning Piler 0.1.23

crontab -l -u piler | grep mpstat

*/5 * * * * LC_ALL=C mpstat | tail -1 | awk '{print }' > /var/piler/stat/cpu.stat

netstat -antl

...
tcp        0      0 0.0.0.0:25              0.0.0.0:*               LISTEN     
...

config-site.php

<?php

$config['SITE_NAME'] = 'mailpiler-test.X.com';
$config['SITE_URL'] = 'http://mailpiler-test.X.com/';
$config['DIR_BASE'] = '/var/www/mailpiler/';

$config['ENABLE_SYSLOG'] = 1;

$config['SMTP_DOMAIN'] = 'mailpiler-test.X.com';
$config['SMTP_FROMADDR'] = 'no-reply@mailpiler-test.X.com';
$config['ADMIN_EMAIL'] = 'admin@mailpiler-test.X.com';

$config['DB_DRIVER'] = 'mysql';
$config['DB_PREFIX'] = ''; 
$config['DB_HOSTNAME'] = 'localhost';
$config['DB_USERNAME'] = 'piler';
$config['DB_PASSWORD'] = Yeah, right!!;
$config['DB_DATABASE'] = 'piler';

$config['SMARTHOST'] = '';
$config['SMARTHOST_PORT'] = 25; 

?>

Thank you!

Emilio.

Comments (15)

  1. Janos SUTO repo owner

    please fix the cronjob to have "print $11" and not only "print"

    In config-site.php, please assign a value to SMARTHOST, and add

    $config['PILERHOST'] = '1.2.3.4';

  2. Emilio Martin Lundgaard Lopez reporter

    Thank you for your quick answer!! The CPU usage got fixed but I still cannot get the SMTP thing right...In both SMARTHOST and PILERHOST I used the private IP I'm using for testing 10.11.2.40 does that make sense or am I missing something?

    Emilio.

  3. Janos SUTO repo owner

    I made a typo, sorry, it's PILER_HOST. The value can be a private address, however you should set them to the same value only if the archive and the smart host are on the same computer. In this case you have to adjust the PILER_PORT and SMARTHOST_PORT

  4. Emilio Martin Lundgaard Lopez reporter

    Hi! Thanks again for your answer...I changed that but still get the error. Everything is on the same computer right now so I should have this, right?

    $config['SMARTHOST'] = '10.11.2.40';
    $config['SMARTHOST_PORT'] = 25; 
    
    $config['PILER_HOST'] = '10.11.2.40';
    $config['PILER_PORT'] = '25';
    

    EMILIO!

  5. Janos SUTO repo owner

    You are almost there. However two programs can't occupy the same IP-address + port. So where does piler listen (address+port)? And where does your mail server listen (address + port)?

  6. Emilio Martin Lundgaard Lopez reporter

    So SMARTHOST is the mail server? If I check netstat it says

    tcp 0 0 0.0.0.0:25 0.0.0.0:* LISTEN 2377/piler

    from which I'm assuming piler is listening in the host's IP and port 25. Can I use 0.0.0.0 in config-site.php or should I use the actual IP?

    Thank you very much for your kind help!

    EMILIO!

  7. Emilio Martin Lundgaard Lopez reporter

    So if I was using Postfix I need to set the smarthost to point to where postfix is install? Sorry if it's a silly question I'm quite a noob in this topic :P

    Emilio.

  8. s s

    Hi, i have the same issue with CPU usage status on health monitor page.

    Here is piler cron:

    crontab -l -u piler

    /5 * * * LC_ALL=C mpstat | tail -1 | awk '{print $11}' > /var/piler/stat/cpu.stat

    cat /var/piler/stat/cpu.stat

    90.88

    piler -V

    piler 0.1.23, build 789, Janos SUTO sj@acts.hu

    Build Date: Fri Aug 16 05:57:45 EDT 2013 ldd version: ldd (GNU libc) 2.12 gcc version: gcc version 4.4.7 20120313 (Red Hat 4.4.7-3) (GCC) Configure command: ./configure --localstatedir=/var --enable-starttls --enable-tcpwrappers

    Is there any other file i have to change? Thanks.

  9. Janos SUTO repo owner

    A ~90 value should be fine, meaning a ~10% load. How much cpu load can you see after refreshing the health page?

  10. s s

    It is 100% all the time at the health page

    And it does not change at all.

    I suppose it should show 10 % in green ?

  11. Janos SUTO repo owner

    yes, it should be ~10% or so. I recommend you to wait a little until the final 0.1.24, and it fixes this issue.

  12. s s

    I fixed it by proper permissions It was:

    ls -ld /var/piler/

    drwx------. 7 piler piler 4096 Aug 21 13:32 /var/piler/

    So just make: chmod 755 /var/piler/

  13. alexminator

    hi i change my config-site.php n put : $config['SMARTHOST'] = '10.122.193.225'; $config['SMARTHOST_PORT'] = 25; $config['PILER_HOST'] = '10.122.193.226'; where 10.122.193.225 its my zimbra(postfix) 10.122.193.226 its IP of Piler but i have in smtp status: piler: ERROR smarthost: OK And if i put in PILER_HOST the ip of my zimbra(postfix) then all its OK PILER_HOST is the ip of piler???or not?? How i fix this...if PILER_HOST its the ip of piler....what i do wrong??

  14. Log in to comment