CiviCRM Community Forums (archive)

I've set up civimail.cronjob.php to run every 5 minutes via the CLI with a nice of -19. The Mailer Batch Limit is 300 emails, and the Mailer job size is 0 (no multi-threading).

Usually one civimail.cronjob.php finishes before the next starts. All is good.

But if the next job starts before the previous one finishes. problems occur. The pattern is this: when 3 are running simultaneously, there is a SASL connection error to the SMTP server, and then a few hundred emails bounce with a "not connected" error. The server also throws out-of-memory errors.

As a workaround, I've put a check into my shell script to exit if another civimail.cronjob.php is already running.

Here's the backtrace ...

Sep 14 12:18:33  [info] $backTrace = /var/www/xxx/sites/all/modules/civicrm/CRM/Core/Error.php, backtrace, 541
, nullHandler,
/var/www/xxx/sites/all/modules/civicrm/packages/PEAR.php, call_user_func, 931
/var/www/xxx/sites/all/modules/civicrm/packages/PEAR.php, PEAR_Error, 566
/var/www/xxx/sites/all/modules/civicrm/packages/Net/Socket.php, raiseError, 327
/var/www/xxx/sites/all/modules/civicrm/packages/Net/SMTP.php, write, 229
/var/www/xxx/sites/all/modules/civicrm/packages/Net/SMTP.php, _send, 265
/var/www/xxx/sites/all/modules/civicrm/packages/Net/SMTP.php, _put, 1138
/var/www/xxx/sites/all/modules/civicrm/packages/Mail/smtp.php, rset, 293
/var/www/xxx/sites/all/modules/civicrm/CRM/Mailing/BAO/Job.php, send, 547
/var/www/xxx/sites/all/modules/civicrm/CRM/Mailing/BAO/Job.php, deliverGroup, 476
/var/www/xxx/sites/all/modules/civicrm/CRM/Mailing/BAO/Job.php, deliver, 152
/var/www/xxx/sites/all/modules/civicrm/bin/civimail.cronjob.php, runJobs, 47
/var/www/xxx/sites/all/modules/civicrm/bin/civimail.cronjob.php, processQueue, 79

The error object is huge: here's the top of it ...

Sep 14 12:18:33  [ info ] $Ignoring exception thrown here = PEAR_Error Object
(
    [ error_message_prefix ] =>
    [ mode ] => 16
    [ level ] => 1024
    [ code ] =>
    [ message ] => Failed to write to socket: not connected
    [ userinfo ] =>
    [ backtrace ] => Array
        (
            [ 0 ] => Array
                (
                    [ file ] => /var/www/xxx/sites/all/modules/civicrm/packages/PEAR.php
                    [ line ] => 566
                    [ function ] => PEAR_Error
                    [ class ] => PEAR_Error
                    [ type ] => ->
                    [ args ] => Array
                        (
                            [ 0 ] => Failed to write to socket: not connected
                            [ 1 ] =>
                            [ 2 ] => 16
                            [ 3 ] => Array
                                (
                                    [ 0 ] => CRM_Core_Error
                                    [ 1 ] => nullHandler
                                )

                            [ 4 ] =>
                        )

                )

Syslog says ...

Sep 14 12:15:21 xxx postfix/smtpd[30583]: warning: localhost[127.0.0.1]: SASL LOGIN authentication failed: Connection lost to authentication server

Note that over 200 emails bounce with status "not connected" between 12:15 (the error in Syslog) and 12:18 (the error in the CiviCRM log). The error object contains an array of 300 emails: I assume what happens is that the "not connect" error occurred before 100 had been sent, and all the subsequent emails failed.

Excellent!

I'm planning on upgrading to 3.4.6 when it's released.

Thanks, Lobo