On Jul 31, 2009, at 10:04 AM, Nathan M <locu.li...@gmail.com> wrote:

> On Fri, Jul 31, 2009 at 4:26 AM, Sahil Tandon<sa...@tandon.net> wrote:
>> I've seen discussion of this before with respect to mysql's  
>> wait_timeout
>> variable, but my question is a little different.
>>
>> If I restart amavisd-new and mysql, mail flows fine.  This is a  
>> test machine
>> so if I try to send another test message the next day (i.e. this  
>> morning), I
>> start seeing this again:
>>
>>  Jul 31 07:13:43 polymath amavis[30861]: (30861-03) NOTICE:  
>> reconnecting in
>>  response to: err=2006, HY000, DBD::mysql::st execute failed: MySQL  
>> server has
>>  gone away at (eval 102) line 166, <GEN21> line 307.
>>
>> I understand why this might happen after a very long time as a new  
>> connection
>> to the mysql server needs to be re-established, but the problem is  
>> that the
>> above line re-appears for *every* new test message thereafter,  
>> which is what
>> puzzles me.  i.e.:
>>
>>  Jul 31 07:14:30 polymath amavis[30862]: (30862-02) NOTICE:  
>> reconnecting in
>>  response to: err=2006, HY000, DBD::mysql::st execute failed: MySQL  
>> server has
>>  gone away at (eval 102) line 166, <GEN21> line 726.
>>
>> Does this occur because different amavisd processes are handling the
>> consecutively new messages and each establish their own connection  
>> to the
>> mysql server?  I haven't debugged that aspect too closely but just  
>> wondering
>> if anyone else has experienced this.  If not then I will have to  
>> debug on the
>> mysql server.  Thanks.
>
> Yup, each process makes it's own mysql connection.  So however many
> processes you have setup to run in amavisd should be the number of
> this error you see.  If you have 10 processes, and send more than 10
> emails through the system, and see that error more than 10 times,
> there might be a prob.  Otherwise, I imagine it's just those idle
> connections timing out on the mysql end and you have nothing to worry
> about as it's all working as expected.

Seems logical hence my theory.  I'll test this scenario in practice  
after the wait_timeout is likely to have expired for all running  
processes.  Thanks for the quick reply Nathan.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
AMaViS-user mailing list
AMaViS-user@lists.sourceforge.net 
https://lists.sourceforge.net/lists/listinfo/amavis-user 
 AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3 
 AMaViS-HowTos:http://www.amavis.org/howto/ 

Reply via email to